maditr: Fast Data Aggregation, Modification, and Filtering

Links

• maditr on CRAN
• maditr on Github
• Issues

Installation

maditr is on CRAN, so for installation you can print in the console install.packages("maditr").

Overview

Package provides pipe-style interface for data.table package. It preserves all data.table features without significant impact on performance. let and take functions are simplified interfaces for most common data manipulation tasks.

• To select rows from data: rows(mtcars, am==0)
• To select columns from data: columns(mtcars, mpg, vs:carb)
• To aggregate data: take(mtcars, mean_mpg = mean(mpg), by = am)
• To aggregate all non-grouping columns: take_all(mtcars, mean, by = am)
• To aggregate several columns with one summary: take(mtcars, mpg, hp, fun = mean, by = am)
• To get total summary skip by argument: take_all(mtcars, mean)
• Use magrittr pipe %>% to chain several operations:

     mtcars %>%
        let(mpg_hp = mpg/hp) %>%
        take(mean(mpg_hp), by = am)

• To modify variables or add new variables:

      mtcars %>%
         let(new_var = 42,
             new_var2 = new_var*hp) %>%
         head()

• To drop variable assign NULL: let(mtcars, am = NULL) %>% head()
• To modify all non-grouping variables:

    iris %>%
      let_all(
          scaled = (.x - mean(.x))/sd(.x),
          by = Species) %>%
       head()

• To aggregate all variables conditionally on name:

    iris %>%
      take_all(
          mean = if(startsWith(.name, "Sepal")) mean(.x),
          median = if(startsWith(.name, "Petal")) median(.x),
          by = Species
      )

• For parametric assignment use :=:

    new_var = "my_var"
    old_var = "mpg"
    mtcars %>%
        let((new_var) := get(old_var)*2) %>%
        head()
     
    # or,  
    expr = quote(mean(cyl))
    mtcars %>% 
        let((new_var) := eval(expr)) %>% 
        head()
    
    # the same with `take` 
    by_var = "vs,am"
    take(mtcars, (new_var) := eval(expr), by = by_var)

query_if function translates its arguments one-to-one to [.data.table method. Additionally there are some conveniences such as automatic data.frame conversion to data.table.

vlookup & xlookup

Let's make datasets for lookups:

library(maditr)

workers = fread("
    name company
    Nick Acme
    John Ajax
    Daniela Ajax
")

positions = fread("
    name position
    John designer
    Daniela engineer
    Cathie manager
")

# xlookup
workers = let(workers,
  position = xlookup(name, positions$name, positions$position)
)

# vlookup
# by default we search in the first column and return values from second column
workers = let(workers,
  position = vlookup(name, positions, no_match = "Not found")
)

# the same 
workers = let(workers,
  position = vlookup(name, positions, 
                     result_column = "position", 
                     no_match = "Not found") # or, result_column = 2 
)

head(workers)

More examples

We will use for demonstartion well-known mtcars dataset and some examples from dplyr package.

library(maditr)

data(mtcars)

# Newly created variables are available immediately
mtcars %>%
    let(
        cyl2 = cyl * 2,
        cyl4 = cyl2 * 2
    ) %>% head()

# You can also use let() to remove variables and
# modify existing variables
mtcars %>%
    let(
        mpg = NULL,
        disp = disp * 0.0163871 # convert to litres
    ) %>% head()


# window functions are useful for grouped computations
mtcars %>%
    let(rank = rank(-mpg, ties.method = "min"),
        by = cyl) %>%
    head()

# You can drop variables by setting them to NULL
mtcars %>%
    let(cyl = NULL) %>%
    head()

# keeps all existing variables
mtcars %>%
    let(displ_l = disp / 61.0237) %>%
    head()

# keeps only the variables you create
mtcars %>%
    take(displ_l = disp / 61.0237) %>% 
    head()


# can refer to both contextual variables and variable names:
var = 100
mtcars %>%
    let(cyl = cyl * var) %>%
    head()

# select rows
mtcars %>%
    rows(am==0) %>% 
    head()

# select rows with compound condition
mtcars %>%
    rows(am==0 & mpg>mean(mpg))

# select columns
mtcars %>% 
    columns(vs:carb, cyl)
    
mtcars %>% 
    columns(-am, -cyl)    

# regular expression pattern
columns(iris, "^Petal") # variables which start from 'Petal'
columns(iris, "Width$") # variables which end with 'Width'

# move Species variable to the front
# pattern "^." matches all variables
columns(iris, Species, "^.")

# pattern "^.*al" means "contains 'al'"
columns(iris, "^.*al")

# numeric indexing - all variables except Species
columns(iris, 1:4) 

# A 'take' with summary functions applied without 'by' argument returns an aggregated data
mtcars %>%
    take(mean = mean(disp), n = .N)

# Usually, you'll want to group first
mtcars %>%
    take(mean = mean(disp), n = .N, by = am)

# grouping by multiple variables
mtcars %>%
    take(mean = mean(disp), n = .N, by = list(am, vs))

# You can group by expressions:
mtcars %>%
    take_all(
        mean,
        by = list(vsam = vs + am)
    )

# modify all non-grouping variables in-place
mtcars %>%
    let_all((.x - mean(.x))/sd(.x), by = am) %>%
    head()

# modify all non-grouping variables to new variables
mtcars %>%
    let_all(scaled = (.x - mean(.x))/sd(.x), by = am) %>%
    head()

# conditionally modify all variables
iris %>%
    let_all(mean = if(is.numeric(.x)) mean(.x)) %>%
    head()

# modify all variables conditionally on name
iris %>%
    let_all(
        mean = if(startsWith(.name, "Sepal")) mean(.x),
        median = if(startsWith(.name, "Petal")) median(.x),
        by = Species
    ) %>%
    head()

# aggregation with 'take_all'
mtcars %>%
    take_all(mean = mean(.x), sd = sd(.x), n = .N, by = am)

# conditionally aggregate all variables
iris %>%
    take_all(mean = if(is.numeric(.x)) mean(.x))

# aggregate all variables conditionally on name
iris %>%
    take_all(
        mean = if(startsWith(.name, "Sepal")) mean(.x),
        median = if(startsWith(.name, "Petal")) median(.x),
        by = Species
    )

# parametric evaluation:
var = quote(mean(cyl))
mtcars %>% 
    let(mean_cyl = eval(var)) %>% 
    head()
take(mtcars, eval(var))

# all together
new_var = "mean_cyl"
mtcars %>% 
    let((new_var) := eval(var)) %>% 
    head()
take(mtcars, (new_var) := eval(var))

Variable selection in the expressions

You can use 'columns' inside expression in the 'take'/'let'. 'columns' will be replaced with data.table with selected columns. In 'let' in the expressions with ':=', 'cols' or '%to%' can be placed in the left part of the expression. It is usefull for multiple assignment. There are four ways of column selection:

1. Simply by column names
2. By variable ranges, e. g. vs:carb. Alternatively, you can use '%to%' instead of colon: 'vs %to% carb'.
3. With regular expressions. Characters which start with '^' or end with $ considered as Perl-style regular expression patterns. For example, '^Petal' returns all variables started with 'Petal'. 'Width$' returns all variables which end with 'Width'. Pattern '^.' matches all variables and pattern '^.*my_str' is equivalent to contains "my_str"'.
4. By character variables with interpolated parts. Expression in the curly brackets inside characters will be evaluated in the parent frame with 'text_expand' function. For example, a{1:3} will be transformed to the names 'a1', 'a2', 'a3'. 'cols' is just a shortcut for 'columns'.

# range selection
iris %>% 
    let(
        avg = rowMeans(Sepal.Length %to% Petal.Width)
    ) %>% 
    head()

# multiassignment
iris %>% 
    let(
        # starts with Sepal or Petal
        multipled1 %to% multipled4 := cols("^(Sepal|Petal)")*2
    ) %>% 
    head()


mtcars %>% 
    let(
        # text expansion
        cols("scaled_{names(mtcars)}") := lapply(cols("{names(mtcars)}"), scale)
    ) %>% 
    head()

# range selection in 'by'
# selection of range + additional column
mtcars %>% 
    take(
        res = sum(cols(mpg, disp %to% drat)),
        by = vs %to% gear
    )

Joins

Here we use the same datasets as with lookups:

workers = fread("
    name company
    Nick Acme
    John Ajax
    Daniela Ajax
")

positions = fread("
    name position
    John designer
    Daniela engineer
    Cathie manager
")

workers
positions

Different kinds of joins:

workers %>% dt_inner_join(positions)
workers %>% dt_left_join(positions)
workers %>% dt_right_join(positions)
workers %>% dt_full_join(positions)

# filtering joins
workers %>% dt_anti_join(positions)
workers %>% dt_semi_join(positions)

To suppress the message, supply by argument:

workers %>% dt_left_join(positions, by = "name")

Use a named by if the join variables have different names:

positions2 = setNames(positions, c("worker", "position")) # rename first column in 'positions'
workers %>% dt_inner_join(positions2, by = c("name" = "worker"))

'dplyr'-like interface for data.table.

There are a small subset of 'dplyr' verbs to work with data.table. Note that there is no group_by verb - use by or keyby argument when needed.

• dt_mutate adds new variables or modify existing variables. If data is data.table then it modifies in-place.
• dt_summarize computes summary statistics. Splits the data into subsets, computes summary statistics for each, and returns the result in the "data.table" form.
• dt_summarize_all the same as dt_summarize but work over all non-grouping variables.
• dt_filter Selects rows/cases where conditions are true. Rows where the condition evaluates to NA are dropped.
• dt_select Selects column/variables from the data set. Range of variables are supported, e. g. vs:carb. Characters which start with ^ or end with \$ considered as Perl-style regular expression patterns. For example, '^Petal' returns all variables started with 'Petal'. 'Width\$' returns all variables which end with 'Width'. Pattern ^. matches all variables and pattern '^.*my_str' is equivalent to contains "my_str". See examples.

# examples from 'dplyr'
# newly created variables are available immediately
mtcars  %>%
    dt_mutate(
        cyl2 = cyl * 2,
        cyl4 = cyl2 * 2
    ) %>%
    head()


# you can also use dt_mutate() to remove variables and
# modify existing variables
mtcars %>%
    dt_mutate(
        mpg = NULL,
        disp = disp * 0.0163871 # convert to litres
    ) %>%
    head()


# window functions are useful for grouped mutates
mtcars %>%
    dt_mutate(
        rank = rank(-mpg, ties.method = "min"),
        keyby = cyl) %>%
    print()


# You can drop variables by setting them to NULL
mtcars %>% dt_mutate(cyl = NULL) %>% head()

# A summary applied without by returns a single row
mtcars %>%
    dt_summarise(mean = mean(disp), n = .N)

# Usually, you'll want to group first
mtcars %>%
    dt_summarise(mean = mean(disp), n = .N, by = cyl)


# Multiple 'by' - variables
mtcars %>%
    dt_summarise(cyl_n = .N, by = list(cyl, vs))

# Newly created summaries immediately
# doesn't overwrite existing variables
mtcars %>%
    dt_summarise(disp = mean(disp),
                  sd = sd(disp),
                  by = cyl)

# You can group by expressions:
mtcars %>%
    dt_summarise_all(mean, by = list(vsam = vs + am))

# filter by condition
mtcars %>%
    dt_filter(am==0)

# filter by compound condition
mtcars %>%
    dt_filter(am==0,  mpg>mean(mpg))


# select
mtcars %>% dt_select(vs:carb, cyl)
mtcars %>% dt_select(-am, -cyl)

# regular expression pattern
dt_select(iris, "^Petal") # variables which start from 'Petal'
dt_select(iris, "Width$") # variables which end with 'Width'
# move Species variable to the front
# pattern "^." matches all variables
dt_select(iris, Species, "^.")
# pattern "^.*al" means "contains 'al'"
dt_select(iris, "^.*al")
dt_select(iris, 1:4) # numeric indexing - all variables except Species