ANALYZE | ANALYSE

Description

ANALYZE collects statistics about ordinary tables in a database and stores the results in the PG_STATISTIC and PG_STATISTIC_EXT system catalogs. After you run the ANALYZE command, you can query the collected statistics in the preceding system catalogs or system views PG_STATS and PG_EXT_STATS. The execution plan generator uses these statistics to generate the most effective execution plan.

If no parameters are specified, ANALYZE analyzes each table and partitioned table in the database. You can also specify the table_name, column, and partition_name parameters to restrict the analysis to a specific table, column, or partitioned table.

ANALYZE | ANALYSE VERIFY checks whether data files of ordinary tables in a database are damaged.

Precautions

• Non-temporary tables cannot be analyzed in an anonymous block, transaction block, function, or stored procedure. Temporary tables in a stored procedure can be analyzed but their statistics updates cannot be rolled back.
• Remote read is not involved in the ANALYZE VERIFY scenario. Therefore, the remote read parameter does not take effect. If the system detects that a page is damaged due to an error in a key system table, the system directly reports an error and does not continue the detection.
• With no table specified, ANALYZE processes all the tables that the current user has permission to analyze in the current database. With tables specified, ANALYZE processes only the specified tables.
• To perform ANALYZE operation to a table, you must be a table owner or a user granted the VACUUM permission on the table. By default, the system administrator has this permission. However, database owners are allowed to analyze all tables in their databases, except shared catalogs. (The restriction for shared catalogs means that a true database-wide analyze operation can only be executed by the system administrator). ANALYZE skips tables on which users do not have permissions.
• ANALYZE does not collect columns for which comparison or equivalent operations cannot be performed, for example, columns of the cursor type.

Syntax

• Collect statistics information about a table.

  { ANALYZE | ANALYSE } [ VERBOSE ]
      [ table_name [ ( column_name [, ...] ) ] ];

• Collect partition statistics information about a partitioned table. This syntax is not supported currently.

  { ANALYZE | ANALYSE } [ VERBOSE ]
      table_name [ ( column_name [, ...] ) ] PARTITION ( patrition_name ) ;

  

  An ordinary partitioned table supports the syntax but not the function of collecting statistics about specified partitions.

Parameters

• VERBOSE

  Enables the display of progress messages.

  

  If VERBOSE is specified, ANALYZE displays the progress information, indicating the table that is being processed. Statistics about tables are also displayed.

• table_name

  Specifies the name (possibly schema-qualified) of a specific table to analyze. If omitted, all regular tables (but not foreign tables) in the current database are analyzed.

  Currently, you can use ANALYZE to collect statistics only from row-store tables.

  Value range: an existing table name.

• column_name, column_1_name, column_2_name

  Specifies the name of a specific column to analyze. All columns are analyzed by default.

  Value range: an existing column name

• partition_name

  Assumes the table is a partitioned table. You can specify partition_name following the keyword PARTITION to analyze the statistics of this table. Currently, ANALYZE can be performed on partitioned tables, but statistics of specified partitions cannot be analyzed.

  Value range: a partition name of a table

• index_name

  Specifies the name of the specific index table to be analyzed (possibly schema-qualified).

  Value range: an existing table name

• FAST|COMPLETE

  The FAST mode verifies the CRC and page header of the table. If the verification fails, an alarm is generated. In COMPLETE mode, the pointer and tuple of the table are parsed and verified.

• CASCADE

  In CASCADE mode, all indexes of the current table are verified.

Examples

-- Create a table.
gaussdb=# CREATE TABLE customer_info
(
WR_RETURNED_DATE_SK       INTEGER                       ,
WR_RETURNED_TIME_SK       INTEGER                       ,
WR_ITEM_SK                INTEGER               NOT NULL,
WR_REFUNDED_CUSTOMER_SK   INTEGER
)
;

-- Create a partitioned table.
gaussdb=# CREATE TABLE customer_par
(
WR_RETURNED_DATE_SK       INTEGER                       ,
WR_RETURNED_TIME_SK       INTEGER                       ,
WR_ITEM_SK                INTEGER               NOT NULL,
WR_REFUNDED_CUSTOMER_SK   INTEGER
)
PARTITION BY RANGE(WR_RETURNED_DATE_SK)
(
PARTITION P1 VALUES LESS THAN(2452275),
PARTITION P2 VALUES LESS THAN(2452640),
PARTITION P3 VALUES LESS THAN(2453000),
PARTITION P4 VALUES LESS THAN(MAXVALUE)
)
ENABLE ROW MOVEMENT;

-- Run the ANALYZE statement to update statistics.
gaussdb=# ANALYZE customer_info;
gaussdb=# ANALYZE customer_par;

-- Run the ANALYZE VERBOSE statement to update statistics and display table information.
gaussdb=# ANALYZE VERBOSE customer_info;
INFO:  ANALYZE INFO : estimate total rows of "customer_info": scanned 0 pages of total 0 pages with 1 retry times, containing 0 live rows and 0 dead rows,  estimated 0 total rows(datanode pid=38661)
INFO:  ANALYZE INFO : "customer_info": scanned 0 of 0 pages, containing 0 live rows and 0 dead rows; 0 rows in sample, 0 estimated total rows(datanode pid=38661)
ANALYZE

If any environment-related fault occurs, check the logs of CN.

-- Delete the tables.
gaussdb=# DROP TABLE customer_info;
gaussdb=# DROP TABLE customer_par;