data-processing/pandas-data-merge-reshape.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Pandas数据合并与重塑\n",
    "- 合并数据集\n",
    "    - 数据库风格的DataFrame合并\n",
    "        - pd.merge(DataFrameObj1, DataFrameObj2)\n",
    "        - 基于两个dataframe中同名column_name作为merge key，类似inner join\n",
    "        - pd.merge(DataFrameObj1, DataFrameObj2, on='column_name')\n",
    "            - 显式地指定column name\n",
    "        - pd.merge(DataFrameObj1, DataFrameObj2, left_on='column_name1', right_on='column_name2')\n",
    "        - pd.merge(DataFrameObj1, DataFrameObj2, how='outer'):outer join\n",
    "        - pd.merge(DataFrameObj1, DataFrameObj2, on='column_name', how='left'):left outer join\n",
    "        - pd.merge(DataFrameObj1, DataFrameObj2, how='inner')\n",
    "        - 列表：Different join types with how argument\n",
    "        - 列表：merge function arguments\n",
    "    - 索引上的合并\n",
    "        - 指定column name和index进行merge\n",
    "            - pd.merge(DataFrameObj1, DataFrameObj2, left_on='column_name', right_index=True)\n",
    "            - pd.merge(DataFrameObj1, DataFrameObj2, left_on='column_name', right_index=True, how='outer')\n",
    "            - pd.merge(DataFrameObj1, DataFrameObj2, left_on=['column_name1', 'column_name2'], right_index=True)\n",
    "        - 指定index和index进行merge\n",
    "            - pd.merge(DataFrameObj1, DataFrameObj2, how='outer', left_index=True, right_index=True)\n",
    "            - 等价于： DataFrameObj1.join(DataFrameObj2, how='outer')\n",
    "        - 多个dataframe间的join\n",
    "            - DataFrameObj1.join([DataFrameObj2, DataFrameObj3])\n",
    "    - axis连接\n",
    "        - concat\n",
    "            - pd.concat([Series1, Series2, Series3]):合并到1列\n",
    "            - pd.concat([Series1, Series2, Series3], axis=1):合并到1行\n",
    "            - pd.concat([Series1, Series2], axis=1, join='inner'):默认是全外连接，这里显式地指定内连接\n",
    "            - pd.concat([Series1, Series2], axis=1, join_axes=[['index1', 'index2', 'index3', 'index4']]):指定`join_axes`\n",
    "            - result = pd.concat([Series1, Series1, Series2], keys=['key1', 'key2', 'key3']):三个level并不存在，所以组成了层次结构\n",
    "                - result.unstack():摊平\n",
    "            - pd.concat([DataFrameObj1, DataFrameObj2], ignore_index=True)\n",
    "    - 合并重叠数据\n",
    "        - np.where(pd.isnull(Series1), Series2, Series1)\n",
    "        - Series2[:-2].combine_first(Series1[2:])\n",
    "- 层次化索引\n",
    "    - 层次化索引介绍\n",
    "        - 设置MultiIndex(`hierarchically-indexed` object, so-called `partial indexing`)\n",
    "        - unstack():把多level index转成单level index\n",
    "        - unstack().stack():再转成多level index\n",
    "    - 重排分级(levels)顺序\n",
    "        - swaplevel('index_name1', 'index_name2')\n",
    "        - sortlevel(index_position)\n",
    "        - swaplevel(index_position0, index_position1).sortlevel(index_position0)\n",
    "    - 根据级别(level)汇总数据\n",
    "        - sum(level='index_name')\n",
    "        - sum(level='index_name', axis=1)\n",
    "    - 使用DataFrame的列进行索引\n",
    "        - set_index()\n",
    "        - reset_index()\n",
    "    - 整型位置索引\n",
    "        - loc (for labels) \n",
    "        - iloc (for integers)\n",
    "- 重塑和轴向旋转\n",
    "    - 重塑层次化索引\n",
    "        - SeriesObj.DataFrameObj.stack():把一个dataframe 压成一个series，即变成层次结构(level)\n",
    "            - SeriesObj.unstack()\n",
    "            - SeriesObj.unstack(0):level number\n",
    "            - SeriesObj.unstack('level_name')\n",
    "        - stack(dropna=False):stack()默认会过滤缺失值\n",
    "    - 将『长格式』旋转(pivot)为『宽格式』\n",
    "        - 旋转：stack() 和 pivot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "u'/Users/zhangjun/Documents/machine-learning-notes/data-processing'"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "# coding:utf-8\n",
    "from pandas import Series, DataFrame\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "%pwd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 合并数据集\n",
    "### 数据库风格的DataFrame合并"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data1</th>\n",
       "      <th>key</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data1 key\n",
       "0      0   b\n",
       "1      1   b\n",
       "2      2   a\n",
       "3      3   c\n",
       "4      4   a\n",
       "5      5   a\n",
       "6      6   b"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1 = pd.DataFrame({'key': ['b', 'b', 'a', 'c', 'a', 'a', 'b'],\n",
    "                    'data1': range(7)})\n",
    "df2 = pd.DataFrame({'key': ['a', 'b', 'd'],\n",
    "                    'data2': range(3)})\n",
    "df1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data2</th>\n",
       "      <th>key</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>d</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data2 key\n",
       "0      0   a\n",
       "1      1   b\n",
       "2      2   d"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is an example of a many-to-one join; the data in df1 has multiple rows labeled a and b, whereas df2 has only one row for each value in the key column. Calling `merge` with these objects we obtain:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data1</th>\n",
       "      <th>key</th>\n",
       "      <th>data2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>6</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data1 key  data2\n",
       "0      0   b      1\n",
       "1      1   b      1\n",
       "2      6   b      1\n",
       "3      2   a      0\n",
       "4      4   a      0\n",
       "5      5   a      0"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(df1, df2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If not specified, merge uses the overlapping column names as the keys. It’s a good practice to specify explicitly, though:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data1</th>\n",
       "      <th>key</th>\n",
       "      <th>data2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>6</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data1 key  data2\n",
       "0      0   b      1\n",
       "1      1   b      1\n",
       "2      6   b      1\n",
       "3      2   a      0\n",
       "4      4   a      0\n",
       "5      5   a      0"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(df1, df2, on='key')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If the column names are different in each object, you can specify them separately:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data1</th>\n",
       "      <th>lkey</th>\n",
       "      <th>data2</th>\n",
       "      <th>rkey</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>6</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data1 lkey  data2 rkey\n",
       "0      0    b      1    b\n",
       "1      1    b      1    b\n",
       "2      6    b      1    b\n",
       "3      2    a      0    a\n",
       "4      4    a      0    a\n",
       "5      5    a      0    a"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df3 = pd.DataFrame({'lkey': ['b', 'b', 'a', 'c', 'a', 'a', 'b'],\n",
    "                    'data1': range(7)})\n",
    "df4 = pd.DataFrame({'rkey': ['a', 'b', 'd'],\n",
    "                    'data2': range(3)})\n",
    "pd.merge(df3, df4, left_on='lkey', right_on='rkey')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By default merge does an `'inner'` join. Other possible options are `'left'`, `'right'`, and `'outer'`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data1</th>\n",
       "      <th>key</th>\n",
       "      <th>data2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>b</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>b</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>6.0</td>\n",
       "      <td>b</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.0</td>\n",
       "      <td>a</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4.0</td>\n",
       "      <td>a</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5.0</td>\n",
       "      <td>a</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>3.0</td>\n",
       "      <td>c</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>NaN</td>\n",
       "      <td>d</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data1 key  data2\n",
       "0    0.0   b    1.0\n",
       "1    1.0   b    1.0\n",
       "2    6.0   b    1.0\n",
       "3    2.0   a    0.0\n",
       "4    4.0   a    0.0\n",
       "5    5.0   a    0.0\n",
       "6    3.0   c    NaN\n",
       "7    NaN   d    2.0"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(df1, df2, how='outer')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Many-to-many merges have well-defined though not necessarily intuitive behavior. Here’s an example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data1</th>\n",
       "      <th>key</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data1 key\n",
       "0      0   b\n",
       "1      1   b\n",
       "2      2   a\n",
       "3      3   c\n",
       "4      4   a\n",
       "5      5   b"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1 = pd.DataFrame({'key': ['b', 'b', 'a', 'c', 'a', 'b'],\n",
    "                    'data1': range(6)})\n",
    "df2 = pd.DataFrame({'key': ['a', 'b', 'a', 'b', 'd'],\n",
    "                    'data2': range(5)})\n",
    "df1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data2</th>\n",
       "      <th>key</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>d</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data2 key\n",
       "0      0   a\n",
       "1      1   b\n",
       "2      2   a\n",
       "3      3   b\n",
       "4      4   d"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data1</th>\n",
       "      <th>key</th>\n",
       "      <th>data2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>b</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>b</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>a</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2</td>\n",
       "      <td>a</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>3</td>\n",
       "      <td>c</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>4</td>\n",
       "      <td>a</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>4</td>\n",
       "      <td>a</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>5</td>\n",
       "      <td>b</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>5</td>\n",
       "      <td>b</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    data1 key  data2\n",
       "0       0   b    1.0\n",
       "1       0   b    3.0\n",
       "2       1   b    1.0\n",
       "3       1   b    3.0\n",
       "4       2   a    0.0\n",
       "5       2   a    2.0\n",
       "6       3   c    NaN\n",
       "7       4   a    0.0\n",
       "8       4   a    2.0\n",
       "9       5   b    1.0\n",
       "10      5   b    3.0"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(df1, df2, on='key', how='left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data1</th>\n",
       "      <th>key</th>\n",
       "      <th>data2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>b</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>b</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>b</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>2</td>\n",
       "      <td>a</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>4</td>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>4</td>\n",
       "      <td>a</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data1 key  data2\n",
       "0      0   b      1\n",
       "1      0   b      3\n",
       "2      1   b      1\n",
       "3      1   b      3\n",
       "4      5   b      1\n",
       "5      5   b      3\n",
       "6      2   a      0\n",
       "7      2   a      2\n",
       "8      4   a      0\n",
       "9      4   a      2"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(df1, df2, how='inner')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To merge with multiple keys, pass a list of column names:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>key1</th>\n",
       "      <th>key2</th>\n",
       "      <th>lval</th>\n",
       "      <th>rval</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>foo</td>\n",
       "      <td>one</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>foo</td>\n",
       "      <td>one</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>foo</td>\n",
       "      <td>two</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bar</td>\n",
       "      <td>one</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>bar</td>\n",
       "      <td>two</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  key1 key2  lval  rval\n",
       "0  foo  one   1.0   4.0\n",
       "1  foo  one   1.0   5.0\n",
       "2  foo  two   2.0   NaN\n",
       "3  bar  one   3.0   6.0\n",
       "4  bar  two   NaN   7.0"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "left = pd.DataFrame({'key1': ['foo', 'foo', 'bar'],\n",
    "                     'key2': ['one', 'two', 'one'],\n",
    "                     'lval': [1, 2, 3]})\n",
    "right = pd.DataFrame({'key1': ['foo', 'foo', 'bar', 'bar'],\n",
    "                      'key2': ['one', 'one', 'one', 'two'],\n",
    "                      'rval': [4, 5, 6, 7]})\n",
    "pd.merge(left, right, on=['key1', 'key2'], how='outer')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`merge` has a `suffixes` option for specifying strings to append to overlapping names in the left and right DataFrame objects:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>key1</th>\n",
       "      <th>key2_x</th>\n",
       "      <th>lval</th>\n",
       "      <th>key2_y</th>\n",
       "      <th>rval</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>foo</td>\n",
       "      <td>one</td>\n",
       "      <td>1</td>\n",
       "      <td>one</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>foo</td>\n",
       "      <td>one</td>\n",
       "      <td>1</td>\n",
       "      <td>one</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>foo</td>\n",
       "      <td>two</td>\n",
       "      <td>2</td>\n",
       "      <td>one</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>foo</td>\n",
       "      <td>two</td>\n",
       "      <td>2</td>\n",
       "      <td>one</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>bar</td>\n",
       "      <td>one</td>\n",
       "      <td>3</td>\n",
       "      <td>one</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>bar</td>\n",
       "      <td>one</td>\n",
       "      <td>3</td>\n",
       "      <td>two</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  key1 key2_x  lval key2_y  rval\n",
       "0  foo    one     1    one     4\n",
       "1  foo    one     1    one     5\n",
       "2  foo    two     2    one     4\n",
       "3  foo    two     2    one     5\n",
       "4  bar    one     3    one     6\n",
       "5  bar    one     3    two     7"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(left, right, on='key1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>key1</th>\n",
       "      <th>key2_left</th>\n",
       "      <th>lval</th>\n",
       "      <th>key2_right</th>\n",
       "      <th>rval</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>foo</td>\n",
       "      <td>one</td>\n",
       "      <td>1</td>\n",
       "      <td>one</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>foo</td>\n",
       "      <td>one</td>\n",
       "      <td>1</td>\n",
       "      <td>one</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>foo</td>\n",
       "      <td>two</td>\n",
       "      <td>2</td>\n",
       "      <td>one</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>foo</td>\n",
       "      <td>two</td>\n",
       "      <td>2</td>\n",
       "      <td>one</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>bar</td>\n",
       "      <td>one</td>\n",
       "      <td>3</td>\n",
       "      <td>one</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>bar</td>\n",
       "      <td>one</td>\n",
       "      <td>3</td>\n",
       "      <td>two</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  key1 key2_left  lval key2_right  rval\n",
       "0  foo       one     1        one     4\n",
       "1  foo       one     1        one     5\n",
       "2  foo       two     2        one     4\n",
       "3  foo       two     2        one     5\n",
       "4  bar       one     3        one     6\n",
       "5  bar       one     3        two     7"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(left, right, on='key1', suffixes=('_left', '_right'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Table 8-1. Different join types with how argument\n",
    "\n",
    "Option | Behavior\n",
    "-------|---------\n",
    "'inner' | Use only the key combinations observed in both tables.\n",
    "'left' | Use all key combinations found in the left table.\n",
    "'right' | Use all key combinations found in the right table.\n",
    "'outer' | Use all key combinations observed in both tables together\n",
    "\n",
    "Table 8-2. merge function arguments\n",
    "\n",
    "Argument | Description\n",
    "---------|------------\n",
    "left | DataFrame to be merged on the left side\n",
    "right | DataFrame to be merged on the right side\n",
    "how | One of 'inner', 'outer', 'left' or 'right'. 'inner' by default\n",
    "on | Column names to join on. Must be found in both DataFrame objects. If not specified and no other join keys given, will use the intersection of the column names in left and right as the join keys\n",
    "left_on | Columns in left DataFrame to use as join keys\n",
    "right_on | Analogous to left_on for left DataFrame\n",
    "left_index | Use row index in left as its join key (or keys, if a MultiIndex)\n",
    "right_index | Analogous to left_index\n",
    "sort | Sort merged data lexicographically by join keys; True by default. Disable to get better performance in some cases on large datasets\n",
    "suffixes | Tuple of string values to append to column names in case of overlap; defaults to ('_x', '_y'). For example, if 'data' in both DataFrame objects, would appear as 'data_x' and 'data_y' in result\n",
    "copy | If False, avoid copying data into resulting data structure in some exceptional cases. By default always copies"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 索引上的合并\n",
    "In some cases, the merge key or keys in a DataFrame will be found in its index. In this case, you can pass `left_index=True` or `right_index=True` (or both) to indicate that the index should be used as the merge key:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>key</th>\n",
       "      <th>value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>a</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>a</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>b</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>c</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  key  value\n",
       "0   a      0\n",
       "1   b      1\n",
       "2   a      2\n",
       "3   a      3\n",
       "4   b      4\n",
       "5   c      5"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "left1 = pd.DataFrame({'key': ['a', 'b', 'a', 'a', 'b', 'c'],\n",
    "                      'value': range(6)})\n",
    "right1 = pd.DataFrame({'group_val': [3.5, 7]}, index=['a', 'b'])\n",
    "left1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>group_val</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   group_val\n",
       "a        3.5\n",
       "b        7.0"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "right1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>key</th>\n",
       "      <th>value</th>\n",
       "      <th>group_val</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>a</td>\n",
       "      <td>2</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>a</td>\n",
       "      <td>3</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>b</td>\n",
       "      <td>4</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  key  value  group_val\n",
       "0   a      0        3.5\n",
       "2   a      2        3.5\n",
       "3   a      3        3.5\n",
       "1   b      1        7.0\n",
       "4   b      4        7.0"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(left1, right1, left_on='key', right_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>key</th>\n",
       "      <th>value</th>\n",
       "      <th>group_val</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>a</td>\n",
       "      <td>2</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>a</td>\n",
       "      <td>3</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>b</td>\n",
       "      <td>4</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>c</td>\n",
       "      <td>5</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  key  value  group_val\n",
       "0   a      0        3.5\n",
       "2   a      2        3.5\n",
       "3   a      3        3.5\n",
       "1   b      1        7.0\n",
       "4   b      4        7.0\n",
       "5   c      5        NaN"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(left1, right1, left_on='key', right_index=True, how='outer')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With hierarchically-indexed data, things are more complicated:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data</th>\n",
       "      <th>key1</th>\n",
       "      <th>key2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.0</td>\n",
       "      <td>Nevada</td>\n",
       "      <td>2001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4.0</td>\n",
       "      <td>Nevada</td>\n",
       "      <td>2002</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data    key1  key2\n",
       "0   0.0    Ohio  2000\n",
       "1   1.0    Ohio  2001\n",
       "2   2.0    Ohio  2002\n",
       "3   3.0  Nevada  2001\n",
       "4   4.0  Nevada  2002"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lefth = pd.DataFrame({'key1': ['Ohio', 'Ohio', 'Ohio', 'Nevada', 'Nevada'],\n",
    "                      'key2': [2000, 2001, 2002, 2001, 2002],\n",
    "                      'data': np.arange(5.)})\n",
    "righth = pd.DataFrame(np.arange(12).reshape((6, 2)),\n",
    "                      index=[['Nevada', 'Nevada', 'Ohio', 'Ohio', 'Ohio', 'Ohio'],\n",
    "                             [2001, 2000, 2000, 2000, 2001, 2002]],\n",
    "                      columns=['event1', 'event2'])\n",
    "lefth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>event1</th>\n",
       "      <th>event2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Nevada</th>\n",
       "      <th>2001</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000</th>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">Ohio</th>\n",
       "      <th>2000</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000</th>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2001</th>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2002</th>\n",
       "      <td>10</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             event1  event2\n",
       "Nevada 2001       0       1\n",
       "       2000       2       3\n",
       "Ohio   2000       4       5\n",
       "       2000       6       7\n",
       "       2001       8       9\n",
       "       2002      10      11"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "righth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data</th>\n",
       "      <th>key1</th>\n",
       "      <th>key2</th>\n",
       "      <th>event1</th>\n",
       "      <th>event2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2000</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2000</td>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2001</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2002</td>\n",
       "      <td>10</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.0</td>\n",
       "      <td>Nevada</td>\n",
       "      <td>2001</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data    key1  key2  event1  event2\n",
       "0   0.0    Ohio  2000       4       5\n",
       "0   0.0    Ohio  2000       6       7\n",
       "1   1.0    Ohio  2001       8       9\n",
       "2   2.0    Ohio  2002      10      11\n",
       "3   3.0  Nevada  2001       0       1"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(lefth, righth, left_on=['key1', 'key2'], right_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data</th>\n",
       "      <th>key1</th>\n",
       "      <th>key2</th>\n",
       "      <th>event1</th>\n",
       "      <th>event2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2001.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2.0</td>\n",
       "      <td>Ohio</td>\n",
       "      <td>2002.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>11.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.0</td>\n",
       "      <td>Nevada</td>\n",
       "      <td>2001.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4.0</td>\n",
       "      <td>Nevada</td>\n",
       "      <td>2002.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NaN</td>\n",
       "      <td>Nevada</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   data    key1    key2  event1  event2\n",
       "0   0.0    Ohio  2000.0     4.0     5.0\n",
       "0   0.0    Ohio  2000.0     6.0     7.0\n",
       "1   1.0    Ohio  2001.0     8.0     9.0\n",
       "2   2.0    Ohio  2002.0    10.0    11.0\n",
       "3   3.0  Nevada  2001.0     0.0     1.0\n",
       "4   4.0  Nevada  2002.0     NaN     NaN\n",
       "4   NaN  Nevada  2000.0     2.0     3.0"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(lefth, righth, left_on=['key1', 'key2'],\n",
    "         right_index=True, how='outer')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using the indexes of both sides of the merge is also possible:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Ohio</th>\n",
       "      <th>Nevada</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Ohio  Nevada\n",
       "a   1.0     2.0\n",
       "c   3.0     4.0\n",
       "e   5.0     6.0"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "left2 = pd.DataFrame([[1., 2.], [3., 4.], [5., 6.]], index=['a', 'c', 'e'],\n",
    "                     columns=['Ohio', 'Nevada'])\n",
    "right2 = pd.DataFrame([[7., 8.], [9., 10.], [11., 12.], [13, 14]],\n",
    "                      index=['b', 'c', 'd', 'e'],\n",
    "                      columns=['Missouri', 'Alabama'])\n",
    "left2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Missouri</th>\n",
       "      <th>Alabama</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>9.0</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>11.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>13.0</td>\n",
       "      <td>14.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Missouri  Alabama\n",
       "b       7.0      8.0\n",
       "c       9.0     10.0\n",
       "d      11.0     12.0\n",
       "e      13.0     14.0"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "right2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Ohio</th>\n",
       "      <th>Nevada</th>\n",
       "      <th>Missouri</th>\n",
       "      <th>Alabama</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>14.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Ohio  Nevada  Missouri  Alabama\n",
       "a   1.0     2.0       NaN      NaN\n",
       "b   NaN     NaN       7.0      8.0\n",
       "c   3.0     4.0       9.0     10.0\n",
       "d   NaN     NaN      11.0     12.0\n",
       "e   5.0     6.0      13.0     14.0"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(left2, right2, how='outer', left_index=True, right_index=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "DataFrame has a convenient `join` instance for merging by index."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Ohio</th>\n",
       "      <th>Nevada</th>\n",
       "      <th>Missouri</th>\n",
       "      <th>Alabama</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>14.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Ohio  Nevada  Missouri  Alabama\n",
       "a   1.0     2.0       NaN      NaN\n",
       "b   NaN     NaN       7.0      8.0\n",
       "c   3.0     4.0       9.0     10.0\n",
       "d   NaN     NaN      11.0     12.0\n",
       "e   5.0     6.0      13.0     14.0"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "left2.join(right2, how='outer')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>key</th>\n",
       "      <th>value</th>\n",
       "      <th>group_val</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>a</td>\n",
       "      <td>0</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>a</td>\n",
       "      <td>2</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>a</td>\n",
       "      <td>3</td>\n",
       "      <td>3.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>b</td>\n",
       "      <td>4</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>c</td>\n",
       "      <td>5</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  key  value  group_val\n",
       "0   a      0        3.5\n",
       "1   b      1        7.0\n",
       "2   a      2        3.5\n",
       "3   a      3        3.5\n",
       "4   b      4        7.0\n",
       "5   c      5        NaN"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "left1.join(right1, on='key')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>New York</th>\n",
       "      <th>Oregon</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>9.0</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>11.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f</th>\n",
       "      <td>16.0</td>\n",
       "      <td>17.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   New York  Oregon\n",
       "a       7.0     8.0\n",
       "c       9.0    10.0\n",
       "e      11.0    12.0\n",
       "f      16.0    17.0"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "another = pd.DataFrame([[7., 8.], [9., 10.], [11., 12.], [16., 17.]],\n",
    "                       index=['a', 'c', 'e', 'f'], columns=['New York', 'Oregon'])\n",
    "another"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Ohio</th>\n",
       "      <th>Nevada</th>\n",
       "      <th>Missouri</th>\n",
       "      <th>Alabama</th>\n",
       "      <th>New York</th>\n",
       "      <th>Oregon</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Ohio  Nevada  Missouri  Alabama  New York  Oregon\n",
       "a   1.0     2.0       NaN      NaN       7.0     8.0\n",
       "c   3.0     4.0       9.0     10.0       9.0    10.0\n",
       "e   5.0     6.0      13.0     14.0      11.0    12.0"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "left2.join([right2, another])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Ohio</th>\n",
       "      <th>Nevada</th>\n",
       "      <th>Missouri</th>\n",
       "      <th>Alabama</th>\n",
       "      <th>New York</th>\n",
       "      <th>Oregon</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>16.0</td>\n",
       "      <td>17.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Ohio  Nevada  Missouri  Alabama  New York  Oregon\n",
       "a   1.0     2.0       NaN      NaN       7.0     8.0\n",
       "b   NaN     NaN       7.0      8.0       NaN     NaN\n",
       "c   3.0     4.0       9.0     10.0       9.0    10.0\n",
       "d   NaN     NaN      11.0     12.0       NaN     NaN\n",
       "e   5.0     6.0      13.0     14.0      11.0    12.0\n",
       "f   NaN     NaN       NaN      NaN      16.0    17.0"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "left2.join([right2, another], how='outer')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### axis连接"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "s1 = pd.Series([0, 1], index=['a', 'b'])\n",
    "s2 = pd.Series([2, 3, 4], index=['c', 'd', 'e'])\n",
    "s3 = pd.Series([5, 6], index=['f', 'g'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Calling `concat` with these object in a list glues together the values and indexes:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a    0\n",
       "b    1\n",
       "c    2\n",
       "d    3\n",
       "e    4\n",
       "f    5\n",
       "g    6\n",
       "dtype: int64"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([s1, s2, s3])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By default concat works along `axis=0`, producing another Series. If you pass `axis=1`, the result will instead be a DataFrame (`axis=1` is the columns):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>g</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     0    1    2\n",
       "a  0.0  NaN  NaN\n",
       "b  1.0  NaN  NaN\n",
       "c  NaN  2.0  NaN\n",
       "d  NaN  3.0  NaN\n",
       "e  NaN  4.0  NaN\n",
       "f  NaN  NaN  5.0\n",
       "g  NaN  NaN  6.0"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([s1, s2, s3], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this case there is no overlap on the other axis, which as you can see is the sorted union (the `'outer' join`) of the indexes. You can instead intersect them by passing `join='inner'`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a    0\n",
       "b    5\n",
       "f    5\n",
       "g    6\n",
       "dtype: int64"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s4 = pd.concat([s1 * 5, s3])\n",
    "s4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>1.0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f</th>\n",
       "      <td>NaN</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>g</th>\n",
       "      <td>NaN</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     0  1\n",
       "a  0.0  0\n",
       "b  1.0  5\n",
       "f  NaN  5\n",
       "g  NaN  6"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([s1, s4], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0  1\n",
       "a  0  0\n",
       "b  1  5"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([s1, s4], axis=1, join='inner')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can even specify the axes to be used on the other axes with `join_axes`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     0    1\n",
       "a  0.0  0.0\n",
       "c  NaN  NaN\n",
       "b  1.0  5.0\n",
       "e  NaN  NaN"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([s1, s4], axis=1, join_axes=[['a', 'c', 'b', 'e']])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A potential issue is that the concatenated pieces are not identifiable in the result. Suppose instead you wanted to create a hierarchical index on the concatenation axis. To do this, use the `keys` argument:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "one    a    0\n",
       "       b    1\n",
       "two    a    0\n",
       "       b    1\n",
       "three  f    5\n",
       "       g    6\n",
       "dtype: int64"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result = pd.concat([s1, s1, s3], keys=['one', 'two', 'three'])\n",
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "      <th>f</th>\n",
       "      <th>g</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>one</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>three</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         a    b    f    g\n",
       "one    0.0  1.0  NaN  NaN\n",
       "two    0.0  1.0  NaN  NaN\n",
       "three  NaN  NaN  5.0  6.0"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result.unstack()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the case of combining Series along `axis=1`, the keys become the DataFrame column headers:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>g</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two  three\n",
       "a  0.0  NaN    NaN\n",
       "b  1.0  NaN    NaN\n",
       "c  NaN  2.0    NaN\n",
       "d  NaN  3.0    NaN\n",
       "e  NaN  4.0    NaN\n",
       "f  NaN  NaN    5.0\n",
       "g  NaN  NaN    6.0"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([s1, s2, s3], axis=1, keys=['one', 'two', 'three'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The same logic extends to DataFrame objects:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  two\n",
       "a    0    1\n",
       "b    2    3\n",
       "c    4    5"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1 = pd.DataFrame(np.arange(6).reshape(3, 2), index=['a', 'b', 'c'],\n",
    "                   columns=['one', 'two'])\n",
    "df2 = pd.DataFrame(5 + np.arange(4).reshape(2, 2), index=['a', 'c'],\n",
    "                   columns=['three', 'four'])\n",
    "df1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>three</th>\n",
       "      <th>four</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   three  four\n",
       "a      5     6\n",
       "c      7     8"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">level1</th>\n",
       "      <th colspan=\"2\" halign=\"left\">level2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>four</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  level1     level2     \n",
       "     one two  three four\n",
       "a      0   1    5.0  6.0\n",
       "b      2   3    NaN  NaN\n",
       "c      4   5    7.0  8.0"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([df1, df2], axis=1, keys=['level1', 'level2'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If you pass a dict of objects instead of a list, the dict’s keys will be used for the `keys` option:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">level1</th>\n",
       "      <th colspan=\"2\" halign=\"left\">level2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>four</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  level1     level2     \n",
       "     one two  three four\n",
       "a      0   1    5.0  6.0\n",
       "b      2   3    NaN  NaN\n",
       "c      4   5    7.0  8.0"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat({'level1': df1, 'level2': df2}, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can name the created axis levels with the `names` argument:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>upper</th>\n",
       "      <th colspan=\"2\" halign=\"left\">level1</th>\n",
       "      <th colspan=\"2\" halign=\"left\">level2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>lower</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "      <th>four</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "upper level1     level2     \n",
       "lower    one two  three four\n",
       "a          0   1    5.0  6.0\n",
       "b          2   3    NaN  NaN\n",
       "c          4   5    7.0  8.0"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([df1, df2], axis=1, keys=['level1', 'level2'],\n",
    "          names=['upper', 'lower'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A last consideration concerns DataFrames in which the row index does not contain any relevant data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.776944</td>\n",
       "      <td>-0.462237</td>\n",
       "      <td>3.057548</td>\n",
       "      <td>-0.332489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2.534262</td>\n",
       "      <td>1.308292</td>\n",
       "      <td>1.561051</td>\n",
       "      <td>0.586242</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.760955</td>\n",
       "      <td>-0.448366</td>\n",
       "      <td>-0.024375</td>\n",
       "      <td>0.974236</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          a         b         c         d\n",
       "0  1.776944 -0.462237  3.057548 -0.332489\n",
       "1  2.534262  1.308292  1.561051  0.586242\n",
       "2  0.760955 -0.448366 -0.024375  0.974236"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1 = pd.DataFrame(np.random.randn(3, 4), columns=['a', 'b', 'c', 'd'])\n",
    "df2 = pd.DataFrame(np.random.randn(2, 3), columns=['b', 'd', 'a'])\n",
    "df1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>b</th>\n",
       "      <th>d</th>\n",
       "      <th>a</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.445224</td>\n",
       "      <td>0.098105</td>\n",
       "      <td>-0.925441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1.758772</td>\n",
       "      <td>-0.565375</td>\n",
       "      <td>-1.338649</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          b         d         a\n",
       "0  1.445224  0.098105 -0.925441\n",
       "1 -1.758772 -0.565375 -1.338649"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this case, you can pass `ignore_index=True`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.776944</td>\n",
       "      <td>-0.462237</td>\n",
       "      <td>3.057548</td>\n",
       "      <td>-0.332489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2.534262</td>\n",
       "      <td>1.308292</td>\n",
       "      <td>1.561051</td>\n",
       "      <td>0.586242</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.760955</td>\n",
       "      <td>-0.448366</td>\n",
       "      <td>-0.024375</td>\n",
       "      <td>0.974236</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.925441</td>\n",
       "      <td>1.445224</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.098105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-1.338649</td>\n",
       "      <td>-1.758772</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.565375</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          a         b         c         d\n",
       "0  1.776944 -0.462237  3.057548 -0.332489\n",
       "1  2.534262  1.308292  1.561051  0.586242\n",
       "2  0.760955 -0.448366 -0.024375  0.974236\n",
       "3 -0.925441  1.445224       NaN  0.098105\n",
       "4 -1.338649 -1.758772       NaN -0.565375"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([df1, df2], ignore_index=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Table 8-3. concat function arguments\n",
    "\n",
    "Argument | Description\n",
    "---------|------------\n",
    "objs | List or dict of pandas objects to be concatenated. The only required argument\n",
    "axis | Axis to concatenate along; defaults to 0\n",
    "join | One of 'inner', 'outer', defaulting to 'outer'; whether to intersection (inner) or union (outer) together indexes along the other axes\n",
    "join_axes | Specific indexes to use for the other n-1 axes instead of performing union/intersection logic\n",
    "keys | Values to associate with objects being concatenated, forming a hierarchical index along the concatenation axis. Can either be a list or array of arbitrary values, an array of tuples, or a list of arrays (if multiple level arrays passed in levels)\n",
    "levels | Specific indexes to use as hierarchical index level or levels if keys passed\n",
    "names | Names for created hierarchical levels if keys and / or levels passed\n",
    "verify_integrity | Check new axis in concatenated object for duplicates and raise exception if so. By default (False) allows duplicates\n",
    "ignore_index | Do not preserve indexes along concatenation axis, instead producing a new range(total_length) index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 合并重叠数据\n",
    "As a motivating example, consider NumPy’s where function, which performs the array-oriented equivalent of an if-else expression:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "f    NaN\n",
       "e    2.5\n",
       "d    NaN\n",
       "c    3.5\n",
       "b    4.5\n",
       "a    NaN\n",
       "dtype: float64"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a = pd.Series([np.nan, 2.5, np.nan, 3.5, 4.5, np.nan],\n",
    "              index=['f', 'e', 'd', 'c', 'b', 'a'])\n",
    "b = pd.Series(np.arange(len(a), dtype=np.float64),\n",
    "              index=['f', 'e', 'd', 'c', 'b', 'a'])\n",
    "a"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "f    0.0\n",
       "e    1.0\n",
       "d    2.0\n",
       "c    3.0\n",
       "b    4.0\n",
       "a    5.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "b"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "f    0.0\n",
       "e    1.0\n",
       "d    2.0\n",
       "c    3.0\n",
       "b    4.0\n",
       "a    NaN\n",
       "dtype: float64"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "b[-1] = np.nan\n",
    "b"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0. ,  2.5,  2. ,  3.5,  4.5,  nan])"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.where(pd.isnull(a), b, a)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Series has a `combine_first` method, which performs the equivalent of this operation along with pandas’s usual data alignment logic:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a    NaN\n",
       "b    4.5\n",
       "c    3.0\n",
       "d    2.0\n",
       "e    1.0\n",
       "f    0.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "b[:-2].combine_first(a[2:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With DataFrames, `combine_first` does the same thing column by column, so you can think of it as “patching” missing data in the calling object with data from the object you pass:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     a    b   c\n",
       "0  1.0  NaN   2\n",
       "1  NaN  2.0   6\n",
       "2  5.0  NaN  10\n",
       "3  NaN  6.0  14"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1 = pd.DataFrame({'a': [1., np.nan, 5., np.nan],\n",
    "                    'b': [np.nan, 2., np.nan, 6.],\n",
    "                    'c': range(2, 18, 4)})\n",
    "df2 = pd.DataFrame({'a': [5., 4., np.nan, 3., 7.],\n",
    "                    'b': [np.nan, 3., 4., 6., 8.]})\n",
    "df1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     a    b\n",
       "0  5.0  NaN\n",
       "1  4.0  3.0\n",
       "2  NaN  4.0\n",
       "3  3.0  6.0\n",
       "4  7.0  8.0"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>14.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     a    b     c\n",
       "0  1.0  NaN   2.0\n",
       "1  4.0  2.0   6.0\n",
       "2  5.0  4.0  10.0\n",
       "3  3.0  6.0  14.0\n",
       "4  7.0  8.0   NaN"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.combine_first(df2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 层次化索引\n",
    "### 层次化索引介绍"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a  1    1.443597\n",
       "   2   -0.909521\n",
       "   3   -0.968799\n",
       "b  1    0.255042\n",
       "   3   -1.125145\n",
       "c  1   -0.691838\n",
       "   2    1.016297\n",
       "d  2    0.833615\n",
       "   3   -0.468632\n",
       "dtype: float64"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.Series(np.random.randn(9),\n",
    "                 index=[['a', 'a', 'a', 'b', 'b', 'c', 'c', 'd', 'd'],\n",
    "                        [1, 2, 3, 1, 3, 1, 2, 2, 3]])\n",
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What you’re seeing is a prettified view of a Series with a `MultiIndex` as its index."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultiIndex(levels=[[u'a', u'b', u'c', u'd'], [1, 2, 3]],\n",
       "           labels=[[0, 0, 0, 1, 1, 2, 2, 3, 3], [0, 1, 2, 0, 2, 0, 1, 1, 2]])"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With a hierarchically-indexed object, so-called `partial indexing` is possible, enabling you to concisely select subsets of the data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    0.255042\n",
       "3   -1.125145\n",
       "dtype: float64"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['b']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "b  1    0.255042\n",
       "   3   -1.125145\n",
       "c  1   -0.691838\n",
       "   2    1.016297\n",
       "dtype: float64"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['b':'c']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "b  1    0.255042\n",
       "   3   -1.125145\n",
       "d  2    0.833615\n",
       "   3   -0.468632\n",
       "dtype: float64"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.loc[['b', 'd']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Selection is even possible in some cases from an “inner” level:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a   -0.909521\n",
       "c    1.016297\n",
       "d    0.833615\n",
       "dtype: float64"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.loc[:, 2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For example, this data could be rearranged into a DataFrame using its `unstack` method:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>1.443597</td>\n",
       "      <td>-0.909521</td>\n",
       "      <td>-0.968799</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>0.255042</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.125145</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>-0.691838</td>\n",
       "      <td>1.016297</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>d</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.833615</td>\n",
       "      <td>-0.468632</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          1         2         3\n",
       "a  1.443597 -0.909521 -0.968799\n",
       "b  0.255042       NaN -1.125145\n",
       "c -0.691838  1.016297       NaN\n",
       "d       NaN  0.833615 -0.468632"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.unstack()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The inverse operation of unstack is `stack`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a  1    1.443597\n",
       "   2   -0.909521\n",
       "   3   -0.968799\n",
       "b  1    0.255042\n",
       "   3   -1.125145\n",
       "c  1   -0.691838\n",
       "   2    1.016297\n",
       "d  2    0.833615\n",
       "   3   -0.468632\n",
       "dtype: float64"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.unstack().stack()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With a DataFrame, either axis can have a hierarchical index:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>Green</th>\n",
       "      <th>Red</th>\n",
       "      <th>Green</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">b</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Ohio     Colorado\n",
       "    Green Red    Green\n",
       "a 1     0   1        2\n",
       "  2     3   4        5\n",
       "b 1     6   7        8\n",
       "  2     9  10       11"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame = pd.DataFrame(np.arange(12).reshape((4, 3)),\n",
    "                     index=[['a', 'a', 'b', 'b'], [1, 2, 1, 2]],\n",
    "                     columns=[['Ohio', 'Ohio', 'Colorado'],\n",
    "                              ['Green', 'Red', 'Green']])\n",
    "frame"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The hierarchical levels can have `names` (as strings or any Python objects). If so, these will show up in the console output (don’t confuse the index names with the axis labels!):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>state</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>color</th>\n",
       "      <th>Green</th>\n",
       "      <th>Red</th>\n",
       "      <th>Green</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>key1</th>\n",
       "      <th>key2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">b</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state      Ohio     Colorado\n",
       "color     Green Red    Green\n",
       "key1 key2                   \n",
       "a    1        0   1        2\n",
       "     2        3   4        5\n",
       "b    1        6   7        8\n",
       "     2        9  10       11"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame.index.names = ['key1', 'key2']\n",
    "frame.columns.names = ['state', 'color']\n",
    "frame"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With partial column indexing you can similarly select groups of columns:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>color</th>\n",
       "      <th>Green</th>\n",
       "      <th>Red</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>key1</th>\n",
       "      <th>key2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">b</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "color      Green  Red\n",
       "key1 key2            \n",
       "a    1         0    1\n",
       "     2         3    4\n",
       "b    1         6    7\n",
       "     2         9   10"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame['Ohio']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A `MultiIndex` can be created by itself and then reused; the columns in the above DataFrame with level names could be created like this:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultiIndex(levels=[[u'Colorado', u'Ohio'], [u'Green', u'Red']],\n",
       "           labels=[[1, 1, 0], [0, 1, 0]],\n",
       "           names=[u'state', u'color'])"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from pandas import MultiIndex\n",
    "MultiIndex.from_arrays([['Ohio', 'Ohio', 'Colorado'], ['Green', 'Red', 'Green']], names=['state', 'color'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 重排分级(levels)顺序\n",
    "The `swaplevel` takes two level numbers or names and returns a new object with the levels interchanged (but the data is otherwise unaltered):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>state</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>color</th>\n",
       "      <th>Green</th>\n",
       "      <th>Red</th>\n",
       "      <th>Green</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>key2</th>\n",
       "      <th>key1</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <th>a</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <th>a</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <th>b</th>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <th>b</th>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state      Ohio     Colorado\n",
       "color     Green Red    Green\n",
       "key2 key1                   \n",
       "1    a        0   1        2\n",
       "2    a        3   4        5\n",
       "1    b        6   7        8\n",
       "2    b        9  10       11"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame.swaplevel('key1', 'key2')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`sortlevel`, on the other hand, sorts the data (stably) using only the values in a single level."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>state</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>color</th>\n",
       "      <th>Green</th>\n",
       "      <th>Red</th>\n",
       "      <th>Green</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>key1</th>\n",
       "      <th>key2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <th>2</th>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state      Ohio     Colorado\n",
       "color     Green Red    Green\n",
       "key1 key2                   \n",
       "a    1        0   1        2\n",
       "b    1        6   7        8\n",
       "a    2        3   4        5\n",
       "b    2        9  10       11"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame.sortlevel(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>state</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>color</th>\n",
       "      <th>Green</th>\n",
       "      <th>Red</th>\n",
       "      <th>Green</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>key2</th>\n",
       "      <th>key1</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">1</th>\n",
       "      <th>a</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">2</th>\n",
       "      <th>a</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state      Ohio     Colorado\n",
       "color     Green Red    Green\n",
       "key2 key1                   \n",
       "1    a        0   1        2\n",
       "     b        6   7        8\n",
       "2    a        3   4        5\n",
       "     b        9  10       11"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame.swaplevel(0, 1).sortlevel(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 根据级别(level)汇总数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>state</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>color</th>\n",
       "      <th>Green</th>\n",
       "      <th>Red</th>\n",
       "      <th>Green</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>key2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>6</td>\n",
       "      <td>8</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12</td>\n",
       "      <td>14</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state  Ohio     Colorado\n",
       "color Green Red    Green\n",
       "key2                    \n",
       "1         6   8       10\n",
       "2        12  14       16"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame.sum(level='key2')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>color</th>\n",
       "      <th>Green</th>\n",
       "      <th>Red</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>key1</th>\n",
       "      <th>key2</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">a</th>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">b</th>\n",
       "      <th>1</th>\n",
       "      <td>14</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>20</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "color      Green  Red\n",
       "key1 key2            \n",
       "a    1         2    1\n",
       "     2         8    4\n",
       "b    1        14    7\n",
       "     2        20   10"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame.sum(level='color', axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 使用DataFrame的列进行索引"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>one</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>one</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>one</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>two</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>two</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>two</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>two</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   a  b    c  d\n",
       "0  0  7  one  0\n",
       "1  1  6  one  1\n",
       "2  2  5  one  2\n",
       "3  3  4  two  0\n",
       "4  4  3  two  1\n",
       "5  5  2  two  2\n",
       "6  6  1  two  3"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame = pd.DataFrame({'a': range(7), 'b': range(7, 0, -1),\n",
    "                      'c': ['one', 'one', 'one', 'two', 'two', 'two', 'two'],\n",
    "                      'd': [0, 1, 2, 0, 1, 2, 3]})\n",
    "frame"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "DataFrame’s `set_index` function will create a new DataFrame using one or more of its columns as the index:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">one</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">two</th>\n",
       "      <th>0</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       a  b\n",
       "c   d      \n",
       "one 0  0  7\n",
       "    1  1  6\n",
       "    2  2  5\n",
       "two 0  3  4\n",
       "    1  4  3\n",
       "    2  5  2\n",
       "    3  6  1"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame2 = frame.set_index(['c', 'd'])\n",
    "frame2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By default the columns are removed from the DataFrame, though you can leave them in:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">one</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>one</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>one</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>one</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">two</th>\n",
       "      <th>0</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>two</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>two</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>two</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>two</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       a  b    c  d\n",
       "c   d              \n",
       "one 0  0  7  one  0\n",
       "    1  1  6  one  1\n",
       "    2  2  5  one  2\n",
       "two 0  3  4  two  0\n",
       "    1  4  3  two  1\n",
       "    2  5  2  two  2\n",
       "    3  6  1  two  3"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame.set_index(['c', 'd'], drop=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`reset_index`, on the other hand, does the opposite of `set_index`; the hierarchical index levels are are moved into the columns:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>one</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>one</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>one</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>two</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>two</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>two</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>two</td>\n",
       "      <td>3</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     c  d  a  b\n",
       "0  one  0  0  7\n",
       "1  one  1  1  6\n",
       "2  one  2  2  5\n",
       "3  two  0  3  4\n",
       "4  two  1  4  3\n",
       "5  two  2  5  2\n",
       "6  two  3  6  1"
      ]
     },
     "execution_count": 117,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame2.reset_index()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 整型位置索引\n",
    "To keep things consistent, if you have an axis index containing integers, data selection will always be label-oriented. For more precise handling, use `loc` (for labels) or `iloc` (for integers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.0\n",
       "1    1.0\n",
       "2    2.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ser = pd.Series(np.arange(3.))\n",
    "ser"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ser[:1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.0\n",
       "1    1.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ser.loc[:1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ser.iloc[:1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 重塑和轴向旋转\n",
    "### 重塑层次化索引"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>number</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>state</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Ohio</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Colorado</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "number    one  two  three\n",
       "state                    \n",
       "Ohio        0    1      2\n",
       "Colorado    3    4      5"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.DataFrame(np.arange(6).reshape((2, 3)),\n",
    "                    index=pd.Index(['Ohio', 'Colorado'], name='state'),\n",
    "                    columns=pd.Index(['one', 'two', 'three'], name='number'))\n",
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using the `stack` method on this data pivots the columns into the rows, producing a Series:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "state     number\n",
       "Ohio      one       0\n",
       "          two       1\n",
       "          three     2\n",
       "Colorado  one       3\n",
       "          two       4\n",
       "          three     5\n",
       "dtype: int64"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result = data.stack()\n",
    "result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From a hierarchically-indexed Series, you can rearrange the data back into a DataFrame with `unstack`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>number</th>\n",
       "      <th>one</th>\n",
       "      <th>two</th>\n",
       "      <th>three</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>state</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Ohio</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Colorado</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "number    one  two  three\n",
       "state                    \n",
       "Ohio        0    1      2\n",
       "Colorado    3    4      5"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result.unstack()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By default the innermost level is unstacked (same with stack). You can unstack a different level by passing a level number or name:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>one</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>three</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state   Ohio  Colorado\n",
       "number                \n",
       "one        0         3\n",
       "two        1         4\n",
       "three      2         5"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result.unstack(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>one</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>three</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state   Ohio  Colorado\n",
       "number                \n",
       "one        0         3\n",
       "two        1         4\n",
       "three      2         5"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result.unstack('state')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Unstacking might introduce missing data if all of the values in the level aren’t found in each of the subgroups:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "a    0\n",
       "b    1\n",
       "c    2\n",
       "d    3\n",
       "dtype: int64"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s1 = pd.Series([0, 1, 2, 3], index=['a', 'b', 'c', 'd'])\n",
    "s2 = pd.Series([4, 5, 6], index=['c', 'd', 'e'])\n",
    "s1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "c    4\n",
       "d    5\n",
       "e    6\n",
       "dtype: int64"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "one  a    0\n",
       "     b    1\n",
       "     c    2\n",
       "     d    3\n",
       "two  c    4\n",
       "     d    5\n",
       "     e    6\n",
       "dtype: int64"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data2 = pd.concat([s1, s2], keys=['one', 'two'])\n",
    "data2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "      <th>e</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>one</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       a    b    c    d    e\n",
       "one  0.0  1.0  2.0  3.0  NaN\n",
       "two  NaN  NaN  4.0  5.0  6.0"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data2.unstack()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Stacking filters out missing data by default, so the operation is more easily invertible:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "one  a    0.0\n",
       "     b    1.0\n",
       "     c    2.0\n",
       "     d    3.0\n",
       "two  c    4.0\n",
       "     d    5.0\n",
       "     e    6.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data2.unstack().stack()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "one  a    0.0\n",
       "     b    1.0\n",
       "     c    2.0\n",
       "     d    3.0\n",
       "     e    NaN\n",
       "two  a    NaN\n",
       "     b    NaN\n",
       "     c    4.0\n",
       "     d    5.0\n",
       "     e    6.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data2.unstack().stack(dropna=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When unstacking in a DataFrame, the level unstacked becomes the lowest level in the result:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>side</th>\n",
       "      <th>left</th>\n",
       "      <th>right</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>state</th>\n",
       "      <th>number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">Ohio</th>\n",
       "      <th>one</th>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>three</th>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">Colorado</th>\n",
       "      <th>one</th>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>4</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>three</th>\n",
       "      <td>5</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "side             left  right\n",
       "state    number             \n",
       "Ohio     one        0      5\n",
       "         two        1      6\n",
       "         three      2      7\n",
       "Colorado one        3      8\n",
       "         two        4      9\n",
       "         three      5     10"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.DataFrame({'left': result, 'right': result + 5},\n",
    "                  columns=pd.Index(['left', 'right'],\n",
    "                                   name='side'))\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>side</th>\n",
       "      <th colspan=\"2\" halign=\"left\">left</th>\n",
       "      <th colspan=\"2\" halign=\"left\">right</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>state</th>\n",
       "      <th>Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "      <th>Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>one</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>two</th>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>6</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>three</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>7</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "side   left          right         \n",
       "state  Ohio Colorado  Ohio Colorado\n",
       "number                             \n",
       "one       0        3     5        8\n",
       "two       1        4     6        9\n",
       "three     2        5     7       10"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.unstack('state')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>state</th>\n",
       "      <th>Ohio</th>\n",
       "      <th>Colorado</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>number</th>\n",
       "      <th>side</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">one</th>\n",
       "      <th>left</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>right</th>\n",
       "      <td>5</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">two</th>\n",
       "      <th>left</th>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>right</th>\n",
       "      <td>6</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">three</th>\n",
       "      <th>left</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>right</th>\n",
       "      <td>7</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state         Ohio  Colorado\n",
       "number side                 \n",
       "one    left      0         3\n",
       "       right     5         8\n",
       "two    left      1         4\n",
       "       right     6         9\n",
       "three  left      2         5\n",
       "       right     7        10"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.unstack('state').stack('side')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 将『长格式』旋转(pivot)为『宽格式』\n",
    "A common way to store multiple time series in databases and CSV is in so-called long or stacked format. FIrst, let’s load some example data and do a small amount of time series wrangling and other data cleaning:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>year</th>\n",
       "      <th>quarter</th>\n",
       "      <th>realgdp</th>\n",
       "      <th>realcons</th>\n",
       "      <th>realinv</th>\n",
       "      <th>realgovt</th>\n",
       "      <th>realdpi</th>\n",
       "      <th>cpi</th>\n",
       "      <th>m1</th>\n",
       "      <th>tbilrate</th>\n",
       "      <th>unemp</th>\n",
       "      <th>pop</th>\n",
       "      <th>infl</th>\n",
       "      <th>realint</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1959.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2710.349</td>\n",
       "      <td>1707.4</td>\n",
       "      <td>286.898</td>\n",
       "      <td>470.045</td>\n",
       "      <td>1886.9</td>\n",
       "      <td>28.98</td>\n",
       "      <td>139.7</td>\n",
       "      <td>2.82</td>\n",
       "      <td>5.8</td>\n",
       "      <td>177.146</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1959.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2778.801</td>\n",
       "      <td>1733.7</td>\n",
       "      <td>310.859</td>\n",
       "      <td>481.301</td>\n",
       "      <td>1919.7</td>\n",
       "      <td>29.15</td>\n",
       "      <td>141.7</td>\n",
       "      <td>3.08</td>\n",
       "      <td>5.1</td>\n",
       "      <td>177.830</td>\n",
       "      <td>2.34</td>\n",
       "      <td>0.74</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1959.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2775.488</td>\n",
       "      <td>1751.8</td>\n",
       "      <td>289.226</td>\n",
       "      <td>491.260</td>\n",
       "      <td>1916.4</td>\n",
       "      <td>29.35</td>\n",
       "      <td>140.5</td>\n",
       "      <td>3.82</td>\n",
       "      <td>5.3</td>\n",
       "      <td>178.657</td>\n",
       "      <td>2.74</td>\n",
       "      <td>1.09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1959.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2785.204</td>\n",
       "      <td>1753.7</td>\n",
       "      <td>299.356</td>\n",
       "      <td>484.052</td>\n",
       "      <td>1931.3</td>\n",
       "      <td>29.37</td>\n",
       "      <td>140.0</td>\n",
       "      <td>4.33</td>\n",
       "      <td>5.6</td>\n",
       "      <td>179.386</td>\n",
       "      <td>0.27</td>\n",
       "      <td>4.06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1960.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2847.699</td>\n",
       "      <td>1770.5</td>\n",
       "      <td>331.722</td>\n",
       "      <td>462.199</td>\n",
       "      <td>1955.5</td>\n",
       "      <td>29.54</td>\n",
       "      <td>139.6</td>\n",
       "      <td>3.50</td>\n",
       "      <td>5.2</td>\n",
       "      <td>180.007</td>\n",
       "      <td>2.31</td>\n",
       "      <td>1.19</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     year  quarter   realgdp  realcons  realinv  realgovt  realdpi    cpi  \\\n",
       "0  1959.0      1.0  2710.349    1707.4  286.898   470.045   1886.9  28.98   \n",
       "1  1959.0      2.0  2778.801    1733.7  310.859   481.301   1919.7  29.15   \n",
       "2  1959.0      3.0  2775.488    1751.8  289.226   491.260   1916.4  29.35   \n",
       "3  1959.0      4.0  2785.204    1753.7  299.356   484.052   1931.3  29.37   \n",
       "4  1960.0      1.0  2847.699    1770.5  331.722   462.199   1955.5  29.54   \n",
       "\n",
       "      m1  tbilrate  unemp      pop  infl  realint  \n",
       "0  139.7      2.82    5.8  177.146  0.00     0.00  \n",
       "1  141.7      3.08    5.1  177.830  2.34     0.74  \n",
       "2  140.5      3.82    5.3  178.657  2.74     1.09  \n",
       "3  140.0      4.33    5.6  179.386  0.27     4.06  \n",
       "4  139.6      3.50    5.2  180.007  2.31     1.19  "
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('data/macrodata.csv')\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PeriodIndex(['1959Q1', '1959Q2', '1959Q3', '1959Q4', '1960Q1', '1960Q2',\n",
       "             '1960Q3', '1960Q4', '1961Q1', '1961Q2',\n",
       "             ...\n",
       "             '2007Q2', '2007Q3', '2007Q4', '2008Q1', '2008Q2', '2008Q3',\n",
       "             '2008Q4', '2009Q1', '2009Q2', '2009Q3'],\n",
       "            dtype='int64', name=u'date', length=203, freq='Q-DEC')"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "periods = pd.PeriodIndex(year=data.year, quarter=data.quarter, name='date')\n",
    "periods"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>item</th>\n",
       "      <th>realgdp</th>\n",
       "      <th>infl</th>\n",
       "      <th>unemp</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1959-03-31</th>\n",
       "      <td>2710.349</td>\n",
       "      <td>0.00</td>\n",
       "      <td>5.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-06-30</th>\n",
       "      <td>2778.801</td>\n",
       "      <td>2.34</td>\n",
       "      <td>5.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-09-30</th>\n",
       "      <td>2775.488</td>\n",
       "      <td>2.74</td>\n",
       "      <td>5.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-12-31</th>\n",
       "      <td>2785.204</td>\n",
       "      <td>0.27</td>\n",
       "      <td>5.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-03-31</th>\n",
       "      <td>2847.699</td>\n",
       "      <td>2.31</td>\n",
       "      <td>5.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "item         realgdp  infl  unemp\n",
       "date                             \n",
       "1959-03-31  2710.349  0.00    5.8\n",
       "1959-06-30  2778.801  2.34    5.1\n",
       "1959-09-30  2775.488  2.74    5.3\n",
       "1959-12-31  2785.204  0.27    5.6\n",
       "1960-03-31  2847.699  2.31    5.2"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.DataFrame(data.to_records(),\n",
    "                    columns=pd.Index(['realgdp', 'infl', 'unemp'], name='item'),\n",
    "                    index=periods.to_timestamp('D', 'end'))\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>item</th>\n",
       "      <th>value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1959-03-31</td>\n",
       "      <td>realgdp</td>\n",
       "      <td>2710.349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1959-03-31</td>\n",
       "      <td>infl</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1959-03-31</td>\n",
       "      <td>unemp</td>\n",
       "      <td>5.800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1959-06-30</td>\n",
       "      <td>realgdp</td>\n",
       "      <td>2778.801</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1959-06-30</td>\n",
       "      <td>infl</td>\n",
       "      <td>2.340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1959-06-30</td>\n",
       "      <td>unemp</td>\n",
       "      <td>5.100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1959-09-30</td>\n",
       "      <td>realgdp</td>\n",
       "      <td>2775.488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1959-09-30</td>\n",
       "      <td>infl</td>\n",
       "      <td>2.740</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1959-09-30</td>\n",
       "      <td>unemp</td>\n",
       "      <td>5.300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1959-12-31</td>\n",
       "      <td>realgdp</td>\n",
       "      <td>2785.204</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        date     item     value\n",
       "0 1959-03-31  realgdp  2710.349\n",
       "1 1959-03-31     infl     0.000\n",
       "2 1959-03-31    unemp     5.800\n",
       "3 1959-06-30  realgdp  2778.801\n",
       "4 1959-06-30     infl     2.340\n",
       "5 1959-06-30    unemp     5.100\n",
       "6 1959-09-30  realgdp  2775.488\n",
       "7 1959-09-30     infl     2.740\n",
       "8 1959-09-30    unemp     5.300\n",
       "9 1959-12-31  realgdp  2785.204"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ldata = data.stack().reset_index().rename(columns={0: 'value'})\n",
    "ldata[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is the so-called `long format` for multiple time series, or other observational data with two or more keys (here, our keys are date and item). Each row in the table represents a single observation.\n",
    "\n",
    "Data is frequently stored this way in relational databases like MySQL as a fixed schema (column names and data types) allows the number of distinct values in the item column to change as data is added to the table. In the above example date and item would usually be the primary keys (in relational database parlance), offering both relational integrity and easier joins. In some cases, the data may be more difficult to work with in this format; you might prefer to have a DataFrame containing one column per distinct item value indexed by timestamps in the date column. DataFrame’s `pivot` method performs exactly this transformation:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>item</th>\n",
       "      <th>infl</th>\n",
       "      <th>realgdp</th>\n",
       "      <th>unemp</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1959-03-31</th>\n",
       "      <td>0.00</td>\n",
       "      <td>2710.349</td>\n",
       "      <td>5.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-06-30</th>\n",
       "      <td>2.34</td>\n",
       "      <td>2778.801</td>\n",
       "      <td>5.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-09-30</th>\n",
       "      <td>2.74</td>\n",
       "      <td>2775.488</td>\n",
       "      <td>5.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-12-31</th>\n",
       "      <td>0.27</td>\n",
       "      <td>2785.204</td>\n",
       "      <td>5.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-03-31</th>\n",
       "      <td>2.31</td>\n",
       "      <td>2847.699</td>\n",
       "      <td>5.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "item        infl   realgdp  unemp\n",
       "date                             \n",
       "1959-03-31  0.00  2710.349    5.8\n",
       "1959-06-30  2.34  2778.801    5.1\n",
       "1959-09-30  2.74  2775.488    5.3\n",
       "1959-12-31  0.27  2785.204    5.6\n",
       "1960-03-31  2.31  2847.699    5.2"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pivoted = ldata.pivot('date', 'item', 'value')\n",
    "pivoted.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first two values passed are the columns to be used respectively as the row and column index, then finally an optional value column to fill the DataFrame. Suppose you had two value columns that you wanted to reshape simultaneously:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>item</th>\n",
       "      <th>value</th>\n",
       "      <th>value2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1959-03-31</td>\n",
       "      <td>realgdp</td>\n",
       "      <td>2710.349</td>\n",
       "      <td>-1.455693</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1959-03-31</td>\n",
       "      <td>infl</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.442000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1959-03-31</td>\n",
       "      <td>unemp</td>\n",
       "      <td>5.800</td>\n",
       "      <td>-0.267403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1959-06-30</td>\n",
       "      <td>realgdp</td>\n",
       "      <td>2778.801</td>\n",
       "      <td>0.220088</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1959-06-30</td>\n",
       "      <td>infl</td>\n",
       "      <td>2.340</td>\n",
       "      <td>-0.339147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1959-06-30</td>\n",
       "      <td>unemp</td>\n",
       "      <td>5.100</td>\n",
       "      <td>0.536942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1959-09-30</td>\n",
       "      <td>realgdp</td>\n",
       "      <td>2775.488</td>\n",
       "      <td>1.689787</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1959-09-30</td>\n",
       "      <td>infl</td>\n",
       "      <td>2.740</td>\n",
       "      <td>-1.254615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1959-09-30</td>\n",
       "      <td>unemp</td>\n",
       "      <td>5.300</td>\n",
       "      <td>-0.562661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1959-12-31</td>\n",
       "      <td>realgdp</td>\n",
       "      <td>2785.204</td>\n",
       "      <td>2.407114</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        date     item     value    value2\n",
       "0 1959-03-31  realgdp  2710.349 -1.455693\n",
       "1 1959-03-31     infl     0.000  0.442000\n",
       "2 1959-03-31    unemp     5.800 -0.267403\n",
       "3 1959-06-30  realgdp  2778.801  0.220088\n",
       "4 1959-06-30     infl     2.340 -0.339147\n",
       "5 1959-06-30    unemp     5.100  0.536942\n",
       "6 1959-09-30  realgdp  2775.488  1.689787\n",
       "7 1959-09-30     infl     2.740 -1.254615\n",
       "8 1959-09-30    unemp     5.300 -0.562661\n",
       "9 1959-12-31  realgdp  2785.204  2.407114"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ldata['value2'] = np.random.randn(len(ldata))\n",
    "ldata[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By omitting the last argument, you obtain a DataFrame with hierarchical columns:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"3\" halign=\"left\">value</th>\n",
       "      <th colspan=\"3\" halign=\"left\">value2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>item</th>\n",
       "      <th>infl</th>\n",
       "      <th>realgdp</th>\n",
       "      <th>unemp</th>\n",
       "      <th>infl</th>\n",
       "      <th>realgdp</th>\n",
       "      <th>unemp</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1959-03-31</th>\n",
       "      <td>0.00</td>\n",
       "      <td>2710.349</td>\n",
       "      <td>5.8</td>\n",
       "      <td>0.442000</td>\n",
       "      <td>-1.455693</td>\n",
       "      <td>-0.267403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-06-30</th>\n",
       "      <td>2.34</td>\n",
       "      <td>2778.801</td>\n",
       "      <td>5.1</td>\n",
       "      <td>-0.339147</td>\n",
       "      <td>0.220088</td>\n",
       "      <td>0.536942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-09-30</th>\n",
       "      <td>2.74</td>\n",
       "      <td>2775.488</td>\n",
       "      <td>5.3</td>\n",
       "      <td>-1.254615</td>\n",
       "      <td>1.689787</td>\n",
       "      <td>-0.562661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-12-31</th>\n",
       "      <td>0.27</td>\n",
       "      <td>2785.204</td>\n",
       "      <td>5.6</td>\n",
       "      <td>-0.172535</td>\n",
       "      <td>2.407114</td>\n",
       "      <td>-0.133247</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-03-31</th>\n",
       "      <td>2.31</td>\n",
       "      <td>2847.699</td>\n",
       "      <td>5.2</td>\n",
       "      <td>1.181717</td>\n",
       "      <td>-0.343066</td>\n",
       "      <td>1.152249</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           value                    value2                    \n",
       "item        infl   realgdp unemp      infl   realgdp     unemp\n",
       "date                                                          \n",
       "1959-03-31  0.00  2710.349   5.8  0.442000 -1.455693 -0.267403\n",
       "1959-06-30  2.34  2778.801   5.1 -0.339147  0.220088  0.536942\n",
       "1959-09-30  2.74  2775.488   5.3 -1.254615  1.689787 -0.562661\n",
       "1959-12-31  0.27  2785.204   5.6 -0.172535  2.407114 -0.133247\n",
       "1960-03-31  2.31  2847.699   5.2  1.181717 -0.343066  1.152249"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pivoted = ldata.pivot('date', 'item')\n",
    "pivoted[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>item</th>\n",
       "      <th>infl</th>\n",
       "      <th>realgdp</th>\n",
       "      <th>unemp</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1959-03-31</th>\n",
       "      <td>0.00</td>\n",
       "      <td>2710.349</td>\n",
       "      <td>5.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-06-30</th>\n",
       "      <td>2.34</td>\n",
       "      <td>2778.801</td>\n",
       "      <td>5.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-09-30</th>\n",
       "      <td>2.74</td>\n",
       "      <td>2775.488</td>\n",
       "      <td>5.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-12-31</th>\n",
       "      <td>0.27</td>\n",
       "      <td>2785.204</td>\n",
       "      <td>5.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-03-31</th>\n",
       "      <td>2.31</td>\n",
       "      <td>2847.699</td>\n",
       "      <td>5.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "item        infl   realgdp  unemp\n",
       "date                             \n",
       "1959-03-31  0.00  2710.349    5.8\n",
       "1959-06-30  2.34  2778.801    5.1\n",
       "1959-09-30  2.74  2775.488    5.3\n",
       "1959-12-31  0.27  2785.204    5.6\n",
       "1960-03-31  2.31  2847.699    5.2"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pivoted['value'][:5]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that pivot is equivalent to creating a hierarchical index using set_index followed by a call to unstack:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"3\" halign=\"left\">value</th>\n",
       "      <th colspan=\"3\" halign=\"left\">value2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>item</th>\n",
       "      <th>infl</th>\n",
       "      <th>realgdp</th>\n",
       "      <th>unemp</th>\n",
       "      <th>infl</th>\n",
       "      <th>realgdp</th>\n",
       "      <th>unemp</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1959-03-31</th>\n",
       "      <td>0.00</td>\n",
       "      <td>2710.349</td>\n",
       "      <td>5.8</td>\n",
       "      <td>0.442000</td>\n",
       "      <td>-1.455693</td>\n",
       "      <td>-0.267403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-06-30</th>\n",
       "      <td>2.34</td>\n",
       "      <td>2778.801</td>\n",
       "      <td>5.1</td>\n",
       "      <td>-0.339147</td>\n",
       "      <td>0.220088</td>\n",
       "      <td>0.536942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-09-30</th>\n",
       "      <td>2.74</td>\n",
       "      <td>2775.488</td>\n",
       "      <td>5.3</td>\n",
       "      <td>-1.254615</td>\n",
       "      <td>1.689787</td>\n",
       "      <td>-0.562661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959-12-31</th>\n",
       "      <td>0.27</td>\n",
       "      <td>2785.204</td>\n",
       "      <td>5.6</td>\n",
       "      <td>-0.172535</td>\n",
       "      <td>2.407114</td>\n",
       "      <td>-0.133247</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-03-31</th>\n",
       "      <td>2.31</td>\n",
       "      <td>2847.699</td>\n",
       "      <td>5.2</td>\n",
       "      <td>1.181717</td>\n",
       "      <td>-0.343066</td>\n",
       "      <td>1.152249</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-06-30</th>\n",
       "      <td>0.14</td>\n",
       "      <td>2834.390</td>\n",
       "      <td>5.2</td>\n",
       "      <td>0.367568</td>\n",
       "      <td>-1.249692</td>\n",
       "      <td>-0.874178</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960-09-30</th>\n",
       "      <td>2.70</td>\n",
       "      <td>2839.022</td>\n",
       "      <td>5.6</td>\n",
       "      <td>-0.824830</td>\n",
       "      <td>-0.063725</td>\n",
       "      <td>1.097307</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           value                    value2                    \n",
       "item        infl   realgdp unemp      infl   realgdp     unemp\n",
       "date                                                          \n",
       "1959-03-31  0.00  2710.349   5.8  0.442000 -1.455693 -0.267403\n",
       "1959-06-30  2.34  2778.801   5.1 -0.339147  0.220088  0.536942\n",
       "1959-09-30  2.74  2775.488   5.3 -1.254615  1.689787 -0.562661\n",
       "1959-12-31  0.27  2785.204   5.6 -0.172535  2.407114 -0.133247\n",
       "1960-03-31  2.31  2847.699   5.2  1.181717 -0.343066  1.152249\n",
       "1960-06-30  0.14  2834.390   5.2  0.367568 -1.249692 -0.874178\n",
       "1960-09-30  2.70  2839.022   5.6 -0.824830 -0.063725  1.097307"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unstacked = ldata.set_index(['date', 'item']).unstack('item')\n",
    "unstacked[:7]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}