Merge branch 'main' into feat/limit_put_memory

CMakeLists.txt

@@ -30,9 +30,14 @@ if (CMAKE_BUILD_TYPE STREQUAL "")
 endif ()
 
 if (NOT DEFINED CMAKE_PREFIX_PATH)
-    set(CMAKE_PREFIX_PATH "${CMAKE_SOURCE_DIR}/.deps/usr")
+    if (DEFINED ENV{THIRD_PARTY_DIR})
+        set(CMAKE_PREFIX_PATH $ENV{THIRD_PARTY_DIR})
+    else()
+        set(CMAKE_PREFIX_PATH "${CMAKE_SOURCE_DIR}/.deps/usr")
+    endif()
 endif()
 
+message (STATUS "CMAKE_PREFIX_PATH: ${CMAKE_PREFIX_PATH}")
 message (STATUS "CMAKE_BUILD_TYPE: ${CMAKE_BUILD_TYPE}")
 set(OPENMLDB_VERSION_MAJOR 0)
 set(OPENMLDB_VERSION_MINOR 8)

docs/en/reference/sql/ddl/SET_STATEMENT.md

@@ -34,6 +34,7 @@ The following format is also equivalent.
 | @@session.enable_trace｜@@enable_trace | When the value is `true`, an error message stack will be printed when the SQL statement has a syntax error or an error occurs during the plan generation process. <br />When the value is `false`, only the basic error message will be printed if there is a SQL syntax error or an error occurs during the plan generation process. | `true`, <br /> `false`      | `false` |
 | @@session.sync_job｜@@sync_job | When the value is `true`, the offline command will be executed synchronously, waiting for the final result of the execution.<br />When the value is `false`, the offline command returns immediately. If you need to check the execution, please use `SHOW JOB` command.                                                              | `true`, <br /> `false`      | `false` |
 | @@session.sync_timeout｜@@sync_timeout | When `sync_job=true`, you can configure the waiting time for synchronization commands. The timeout will return immediately. After the timeout returns, you can still view the command execution through `SHOW JOB`.                                                                                                                   | Int                         | 20000 |
+| @@session.spark_config｜@@spark_config | Set the Spark configuration for offline jobs, configure like 'spark.executor.memory=2g;spark.executor.cores=2'. Notice that the priority of this Spark configuration is higer than TaskManager Spark configuration but lower than CLI Spark configuration file.                                                                                                                    | String                         | "" |
 
 ## Example
 

docs/zh/faq/client_faq.md

@@ -76,13 +76,21 @@ sdk日志（glog日志）：
 
 ## 离线命令Spark报错
 
-`java.lang.OutOfMemoryError: Java heap space`
-
-离线命令的Spark配置默认为`local[*]`，并发较高可能出现OutOfMemoryError错误，请调整`spark.driver.memory`和`spark.executor.memory`两个spark配置项。可以写在TaskManager运行目录的`conf/taskmanager.properties`的`spark.default.conf`并重启TaskManager，或者使用CLI客户端进行配置，参考[客户端Spark配置文件](../reference/client_config/client_spark_config.md)。
 ```
-spark.default.conf=spark.driver.memory=16g;spark.executor.memory=16g
+java.lang.OutOfMemoryError: Java heap space
 ```
 
+```
 Container killed by YARN for exceeding memory limits. 5 GB of 5 GB physical memory used. Consider boosting spark.yarn.executor.memoryOverhead.
+```
+
+出现以上几种日志时，说明离线任务所需资源多于当前配置。一般是这几种情况：
+- 离线命令的Spark配置`local[*]`，机器核数较多，并发度很高，资源占用过大
+- memory配置较小
+
+如果是local模式，单机资源比较有限，可以考虑降低并发度。如果不降低并发，请调整`spark.driver.memory`和`spark.executor.memory`两个spark配置项。可以写在TaskManager运行目录的`conf/taskmanager.properties`的`spark.default.conf`并重启TaskManager，或者使用CLI客户端进行配置，参考[客户端Spark配置文件](../reference/client_config/client_spark_config.md)。
+```
+spark.default.conf=spark.driver.memory=16g;spark.executor.memory=16g
+```
 
-local时drivermemory
+master为local时，不是调整executor的，而是driver的memory，如果你不确定，可以两者都调节。

docs/zh/openmldb_sql/ddl/SET_STATEMENT.md

@@ -35,7 +35,7 @@ sessionVariableName ::= '@@'Identifier | '@@session.'Identifier | '@@global.'Ide
 | @@session.enable_trace｜@@enable_trace | 当该变量值为 `true`，SQL语句有语法错误或者在计划生成过程发生错误时，会打印错误信息栈。<br />当该变量值为 `false`，SQL语句有语法错误或者在计划生成过程发生错误时，仅打印基本错误信息。      | "true" \| "false"     | "false"   |
 | @@session.sync_job｜@@sync_job | 当该变量值为 `true`，离线的命令将变为同步，等待执行的最终结果。<br />当该变量值为 `false`，离线的命令即时返回，若要查看命令的执行情况，请使用`SHOW JOB`。                  | "true" \| "false"     | "false"   |
 | @@session.job_timeout｜@@job_timeout | 可配置离线异步命令或离线管理命令的等待时间（以*毫秒*为单位），将立即返回。离线异步命令返回后仍可通过`SHOW JOB`查看命令执行情况。                             | Int | "20000" |
-
+| @@session.spark_config｜@@spark_config | 设置离线任务的 Spark 参数，配置项参考 'spark.executor.memory=2g;spark.executor.cores=2'。注意此 Spark 配置优先级高于 TaskManager 默认 Spark 配置，低于命令行的 Spark 配置文件。                                                                                                                   | String                         | "" |
 ## Example
 
 ### 设置和显示会话系统变量

docs/zh/openmldb_sql/dml/DELETE_STATEMENT.md

@@ -12,7 +12,7 @@ TableName ::=
 
 **说明**
 
-- `DELETE` 语句删除在线表满足指定条件的数据
+- `DELETE` 语句删除在线表满足指定条件的数据，删除并不是所有索引中满足条件的数据都被删除，只会删除与where condition相关的索引，示例见[功能边界](../../quickstart/function_boundary.md#delete)。
 - `WHERE` 指定的筛选列必须是索引列。如果是key列只能用等于
 
 ## Examples
@@ -25,4 +25,4 @@ DELETE FROM t1 WHERE col1 = 'aaaa' and ts_col = 1687145994000;
 DELETE FROM t1 WHERE col1 = 'aaaa' and ts_col > 1687059594000 and ts_col < 1687145994000;
 
 DELETE FROM t1 WHERE ts_col > 1687059594000 and ts_col < 1687145994000;
-```
+```

docs/zh/openmldb_sql/dml/LOAD_DATA_STATEMENT.md

@@ -1,6 +1,10 @@
 # LOAD DATA INFILE
 `LOAD DATA INFILE`语句能高效地将文件中的数据读取到数据库中的表中。`LOAD DATA INFILE` 与 `SELECT INTO OUTFILE`互补。要将数据从 table导出到文件，请使用[SELECT INTO OUTFILE](../dql/SELECT_INTO_STATEMENT.md)。要将文件数据导入到 table 中，请使用`LOAD DATA INFILE`。
 
+```{note}
+INFILE 的 filePath，既可以是单个文件名，也可以是目录，也可以使用`*`通配符。如果目录中存在多格式的文件，只会选择 LoadDataInfileOptionsList 中指定的FORMAT格式文件。具体格式等价于`DataFrameReader.read.load(String)`，可以使用spark shell来read你想要的文件路径，确认能否读入成功。
+```
+
 ## Syntax
 
 ```sql
@@ -70,7 +74,7 @@ FilePathPattern
 ## SQL语句模版
 
 ```sql
-LOAD DATA INFILE 'file_name' INTO TABLE 'table_name' OPTIONS (key = value, ...);
+LOAD DATA INFILE 'file_path' INTO TABLE 'table_name' OPTIONS (key = value, ...);
 ```
 
 ## Hive 支持

docs/zh/openmldb_sql/dql/GROUP_BY_CLAUSE.md

@@ -2,9 +2,9 @@
 
 ## Syntax
 
-```SQL
-GroupByClause
-         ::= 'GROUP' 'BY' ByList
+```yacc
+group_by_clause:
+  GROUP BY group_by_specification
 ```
 
 ## SQL语句模版

docs/zh/openmldb_sql/dql/HAVING_CLAUSE.md

@@ -4,9 +4,9 @@ Having 子句与 Where 子句作用类似.Having 子句过滤 GroupBy 后的各
 
 ## Syntax
 
-```
-HavingClause
-         ::= 'HAVING' Expression 
+```yacc
+having_clause
+  HAVING bool_expression
 ```
 
 ## SQL语句模版

docs/zh/openmldb_sql/dql/JOIN_CLAUSE.md

@@ -16,16 +16,16 @@ LAST JOIN 是 OpenMLDB SQL 拓展的 JOIN类型. 它的语法和 LEFT JOIN 基
 
 ## Syntax
 
-```
-join:
-    TableRef "LAST" "JOIN" TableRef [OrderByClause] "ON" Expression 
-    | TableRef join_type "JOIN" TableRef "ON" Expression
+```yacc
+join_operation:
+  condition_join_operation
 
-join_type:
-    'LEFT' [OUTER]
+condition_join_operation:
+  from_item LEFT [ OUTER ] JOIN from_item join_condition
+  | from_item LAST JOIN [ ORDER BY ordering_expression ] from_item join_condition
 
-order_by_clause:
-    'ORDER' 'BY' <COLUMN_NAME>
+join_condition:
+  ON bool_expression
 ```
 
 ### 使用限制说明

docs/zh/openmldb_sql/dql/LIMIT_CLAUSE.md

@@ -4,9 +4,9 @@ Limit子句用于限制返回的结果条数。Limit支持接受一个参数，
 
 ## Syntax
 
-```sql
-LimitClause
-         ::= 'LIMIT' int_leteral
+```yacc
+limit_clause:
+  LIMIT numeric_expression
 ```
 
 ## SQL语句模版

docs/zh/openmldb_sql/dql/NO_TABLE_SELECT_CLAUSE.md

@@ -4,14 +4,8 @@
 
 ## Syntax
 
-```sql
-NoTableSelectClause
-	::= 'SELECT' SelectExprList
-SelectExprList
-         ::= SelectExpr ( ',' SelectExpr )*
-SelectExpr    ::= ( Identifier '.' ( Identifier '.' )? )? '*'
-           | ( Expression | '{' Identifier Expression '}' ) ['AS' Identifier]
-
+```yacc
+SELECT { expression [ [ AS ] alias ] } [, ...];
 ```
 
 ## SQL语句模版

docs/zh/openmldb_sql/dql/SELECT_INTO_STATEMENT.md

@@ -5,21 +5,14 @@
 ```
 ## Syntax
 
-```sql
-SelectIntoStmt
-						::= SelectStmt 'INTO' 'OUTFILE' filePath SelectIntoOptionList
-
-filePath 
-						::= string_literal
-SelectIntoOptionList
-						::= 'OPTIONS' '(' SelectInfoOptionItem (',' SelectInfoOptionItem)* ')'
-
-SelectInfoOptionItem
-						::= 'DELIMITER' '=' string_literal
-						|'HEADER' '=' bool_literal
-						|'NULL_VALUE' '=' string_literal
-						|'FORMAT' '=' string_literal
-						|'MODE' '=' string_literal
+```yacc
+select_into_statement:
+  query INTO OUTFILE string_file_path
+  [ OPTIONS options_list ]
+  [ CONFIG options_list ]
+
+options_list:
+  ( { key = value } [, ...] )
 ```
 
 `SELECT INTO OUTFILE`分为三个部分。