On this page 
Use the CREATE STATISTICS statement to generate table statistics for the cost-based optimizer to use.
Once you create a table and load data into it (e.g., INSERT, IMPORT INTO), table statistics can be generated. Table statistics help the cost-based optimizer determine the cardinality of the rows used in each query, which helps to predict more accurate costs.
For compatibility with PostgreSQL, CockroachDB supports the ANALYZE/ANALYSE statement as an alias for CREATE STATISTICS. For syntax, see Aliases.
By default, CockroachDB automatically generates statistics on all indexed columns and up to 100 non-indexed columns, and automatically collects multi-column statistics on columns that prefix each index. As a result, most users do not need to directly issue CREATE STATISTICS statements.
Syntax
Parameters
| Parameter | Description |
|---|---|
statistics_name |
The name of the set of statistics you are creating. |
opt_stats_columns |
The name of the column(s) to create statistics for. |
create_stats_target |
The name of the table to create statistics for. |
opt_as_of_clause |
Create historical stats using the AS OF SYSTEM TIME clause. For instructions, see Create statistics as of a given time. |
Required privileges
The user must have the CREATE privilege on the parent database.
Aliases
For PostgreSQL compatibility, CockroachDB supports ANALYZE and ANALYSE as aliases for CREATE STATISTICS.
Alias syntax
Alias parameters
| Parameter | Description |
|---|---|
analyze_target |
The name of the table to create statistics for. |
Examples
Setup
To follow along, run cockroach demo to start a temporary, in-memory cluster with the movr sample dataset preloaded:
$ cockroach demo
Create statistics on a single column
> CREATE STATISTICS revenue_stats ON revenue FROM rides;
> SHOW STATISTICS FOR TABLE rides;
statistics_name | column_names | created | row_count | distinct_count | null_count | avg_size | histogram_id
------------------+---------------------------+----------------------------+-----------+----------------+------------+----------+---------------------
__auto__ | {city} | 2022-04-20 22:43:08.851613 | 500 | 9 | 0 | 12 | 755053982033936385
__auto__ | {id} | 2022-04-20 22:43:08.851613 | 500 | 500 | 0 | 26 | 755053982039703553
__auto__ | {city,id} | 2022-04-20 22:43:08.851613 | 500 | 500 | 0 | 37 | NULL
__auto__ | {rider_id} | 2022-04-20 22:43:08.851613 | 500 | 50 | 0 | 17 | 755053982050910209
__auto__ | {city,rider_id} | 2022-04-20 22:43:08.851613 | 500 | 50 | 0 | 29 | NULL
__auto__ | {vehicle_city} | 2022-04-20 22:43:08.851613 | 500 | 9 | 0 | 11 | 755053982061690881
__auto__ | {vehicle_id} | 2022-04-20 22:43:08.851613 | 500 | 15 | 0 | 17 | 755053982067392513
__auto__ | {vehicle_city,vehicle_id} | 2022-04-20 22:43:08.851613 | 500 | 15 | 0 | 28 | NULL
__auto__ | {start_address} | 2022-04-20 22:43:08.851613 | 500 | 500 | 0 | 25 | 755053982080991233
__auto__ | {end_address} | 2022-04-20 22:43:08.851613 | 500 | 500 | 0 | 25 | 755053982087544833
__auto__ | {start_time} | 2022-04-20 22:43:08.851613 | 500 | 30 | 0 | 7 | 755053982093443073
__auto__ | {end_time} | 2022-04-20 22:43:08.851613 | 500 | 367 | 0 | 7 | 755053982099472385
__auto__ | {revenue} | 2022-04-20 22:43:08.851613 | 500 | 100 | 0 | 6 | 755053982105337857
revenue_stats | {revenue} | 2022-04-20 22:35:42.279266 | 500 | 100 | 0 | 6 | 755052518729449473
(14 rows)
Statistics are automatically collected for all columns, making the revenue_stats statistics a duplicate of the statistics automatically collected on the revenue column.
Create statistics on multiple columns
> CREATE STATISTICS city_revenue_stats ON city, revenue FROM rides;
> SHOW STATISTICS FOR TABLE rides;
statistics_name | column_names | created | row_count | distinct_count | null_count | avg_size | histogram_id
---------------------+---------------------------+----------------------------+-----------+----------------+------------+----------+---------------------
__auto__ | {city} | 2022-04-20 22:43:08.851613 | 500 | 9 | 0 | 12 | 755053982033936385
__auto__ | {id} | 2022-04-20 22:43:08.851613 | 500 | 500 | 0 | 26 | 755053982039703553
__auto__ | {city,id} | 2022-04-20 22:43:08.851613 | 500 | 500 | 0 | 37 | NULL
__auto__ | {rider_id} | 2022-04-20 22:43:08.851613 | 500 | 50 | 0 | 17 | 755053982050910209
__auto__ | {city,rider_id} | 2022-04-20 22:43:08.851613 | 500 | 50 | 0 | 29 | NULL
__auto__ | {vehicle_city} | 2022-04-20 22:43:08.851613 | 500 | 9 | 0 | 11 | 755053982061690881
__auto__ | {vehicle_id} | 2022-04-20 22:43:08.851613 | 500 | 15 | 0 | 17 | 755053982067392513
__auto__ | {vehicle_city,vehicle_id} | 2022-04-20 22:43:08.851613 | 500 | 15 | 0 | 28 | NULL
__auto__ | {start_address} | 2022-04-20 22:43:08.851613 | 500 | 500 | 0 | 25 | 755053982080991233
__auto__ | {end_address} | 2022-04-20 22:43:08.851613 | 500 | 500 | 0 | 25 | 755053982087544833
__auto__ | {start_time} | 2022-04-20 22:43:08.851613 | 500 | 30 | 0 | 7 | 755053982093443073
__auto__ | {end_time} | 2022-04-20 22:43:08.851613 | 500 | 367 | 0 | 7 | 755053982099472385
__auto__ | {revenue} | 2022-04-20 22:43:08.851613 | 500 | 100 | 0 | 6 | 755053982105337857
revenue_stats | {revenue} | 2022-04-20 22:45:40.272665 | 500 | 100 | 0 | 6 | 755054478211481601
city_revenue_stats | {city,revenue} | 2022-04-20 22:45:46.925799 | 500 | 372 | 0 | 18 | NULL
(15 rows)
Multi-column statistics are automatically collected for all columns that prefix an index. In this example, city and revenue are not an index prefix, making the city_revenue_stats statistics unique for the table.
Create statistics on a default set of columns
The CREATE STATISTICS statement shown below automatically figures out which columns to get statistics on.
> CREATE STATISTICS users_stats FROM users;
This statement creates statistics identical to the statistics that CockroachDB creates automatically.
> SHOW STATISTICS FOR TABLE users;
statistics_name | column_names | created | row_count | distinct_count | null_count | avg_size | histogram_id
------------------+---------------+----------------------------+-----------+----------------+------------+----------+---------------------
__auto__ | {city} | 2022-04-20 22:43:08.819069 | 50 | 9 | 0 | 12 | 755053981927636993
__auto__ | {id} | 2022-04-20 22:43:08.819069 | 50 | 50 | 0 | 27 | 755053981933273089
__auto__ | {city,id} | 2022-04-20 22:43:08.819069 | 50 | 50 | 0 | 38 | NULL
__auto__ | {name} | 2022-04-20 22:43:08.819069 | 50 | 49 | 0 | 16 | 755053981944184833
__auto__ | {address} | 2022-04-20 22:43:08.819069 | 50 | 50 | 0 | 26 | 755053981949001729
__auto__ | {credit_card} | 2022-04-20 22:43:08.819069 | 50 | 50 | 0 | 12 | 755053981954015233
users_stats | {city} | 2022-04-20 22:46:14.878975 | 50 | 9 | 0 | 12 | 755054591614943233
users_stats | {id} | 2022-04-20 22:46:14.878975 | 50 | 50 | 0 | 27 | 755054591622447105
users_stats | {city,id} | 2022-04-20 22:46:14.878975 | 50 | 50 | 0 | 38 | NULL
users_stats | {name} | 2022-04-20 22:46:14.878975 | 50 | 49 | 0 | 16 | 755054591638634497
users_stats | {address} | 2022-04-20 22:46:14.878975 | 50 | 50 | 0 | 26 | 755054591645712385
users_stats | {credit_card} | 2022-04-20 22:46:14.878975 | 50 | 50 | 0 | 12 | 755054591652691969
(12 rows)
Create statistics as of a given time
To create statistics as of a given time (in this example, 1 minute ago to avoid interfering with the production workload), run a statement like the following:
> CREATE STATISTICS vehicle_stats_1 FROM vehicles AS OF SYSTEM TIME '-1m';
For more information about how the AS OF SYSTEM TIME clause works, including supported time formats, see AS OF SYSTEM TIME.
Delete statistics
To delete statistics for all tables in all databases:
DELETE FROM system.table_statistics WHERE true;
To delete a named set of statistics (e.g, one named "users_stats"), run a query like the following:
DELETE FROM system.table_statistics WHERE name = 'users_stats';
For more information about the DELETE statement, see DELETE.
View statistics jobs
Every time the CREATE STATISTICS statement is executed, it starts a background job. This is true for queries issued by your application as well as queries issued for automatically generated statistics.
To view statistics jobs, there are two options:
Use
SHOW JOBSto see all statistics jobs that were created by user queries (i.e., someone enteringCREATE STATISTICSat the SQL prompt or via application code):> WITH x AS (SHOW JOBS) SELECT * FROM x WHERE job_type LIKE '%CREATE STATS%';job_id | job_type | description | statement | user_name | status | running_status | created | started | finished | modified | fraction_completed | error | coordinator_id | trace_id | last_run | next_run | num_runs | execution_errors ---------------------+--------------+--------------------------------------------------------------------------------------------------+-----------+-----------+-----------+----------------+----------------------------+----------------------------+----------------------------+----------------------------+--------------------+-------+----------------+---------------------+----------------------------+----------------------------+----------+------------------- 755053959919108097 | CREATE STATS | CREATE STATISTICS revenue_stats ON revenue FROM movr.public.rides | | demo | succeeded | NULL | 2022-04-20 22:43:02.104229 | 2022-04-20 22:43:02.109754 | 2022-04-20 22:43:02.123381 | 2022-04-20 22:43:02.122569 | 1 | | 1 | 2935658651779633570 | 2022-04-20 22:43:02.109755 | 2022-04-20 22:43:32.109755 | 1 | {} 755054478158364673 | CREATE STATS | CREATE STATISTICS revenue_stats ON revenue FROM movr.public.rides | | demo | succeeded | NULL | 2022-04-20 22:45:40.25829 | 2022-04-20 22:45:40.26361 | 2022-04-20 22:45:40.275882 | 2022-04-20 22:45:40.275032 | 1 | | 1 | 3941365223642966402 | 2022-04-20 22:45:40.263611 | 2022-04-20 22:46:10.263611 | 1 | {} 755054499947053057 | CREATE STATS | CREATE STATISTICS city_revenue_stats ON city, revenue FROM movr.public.rides | | demo | succeeded | NULL | 2022-04-20 22:45:46.907672 | 2022-04-20 22:45:46.912906 | 2022-04-20 22:45:46.929632 | 2022-04-20 22:45:46.928409 | 1 | | 1 | 6666823949040131150 | 2022-04-20 22:45:46.912906 | 2022-04-20 22:46:16.912906 | 1 | {} 755054591559172097 | CREATE STATS | CREATE STATISTICS users_stats FROM movr.public.users | | demo | succeeded | NULL | 2022-04-20 22:46:14.865479 | 2022-04-20 22:46:14.869725 | 2022-04-20 22:46:14.895666 | 2022-04-20 22:46:14.89469 | 1 | | 1 | 6731618360724088456 | 2022-04-20 22:46:14.869726 | 2022-04-20 22:46:44.869726 | 1 | {} 755055101600661505 | CREATE STATS | CREATE STATISTICS vehicle_stats_1 FROM movr.public.vehicles WITH OPTIONS AS OF SYSTEM TIME '-1m' | | demo | succeeded | NULL | 2022-04-20 22:48:50.517787 | 2022-04-20 22:48:50.523002 | 2022-04-20 22:48:50.552397 | 2022-04-20 22:48:50.551333 | 1 | | 1 | 8797094911788373312 | 2022-04-20 22:48:50.523003 | 2022-04-20 22:49:20.523003 | 1 | {} (5 rows)Use
SHOW AUTOMATIC JOBSto see statistics jobs that were created by automatically generated statistics:> WITH x AS (SHOW AUTOMATIC JOBS) SELECT * FROM x WHERE job_type LIKE '%CREATE STATS%';job_id | job_type | description | statement | user_name | status | running_status | created | started | finished | modified | fraction_completed | error | coordinator_id | trace_id | last_run | next_run | num_runs | execution_errors ---------------------+-------------------+---------------------------------------------------------------------+--------------------------------------------------------------------------------------------+-----------+-----------+----------------+----------------------------+----------------------------+----------------------------+----------------------------+--------------------+-------+----------------+---------------------+----------------------------+----------------------------+----------+------------------- 755046859132338177 | AUTO CREATE STATS | Table statistics refresh for movr.public.vehicle_location_histories | CREATE STATISTICS __auto__ FROM [109] WITH OPTIONS THROTTLING 0.9 AS OF SYSTEM TIME '-30s' | root | succeeded | NULL | 2022-04-20 22:06:55.116073 | 2022-04-20 22:06:55.126215 | 2022-04-20 22:06:55.81112 | 2022-04-20 22:06:55.809757 | 1 | | 1 | 7300707583932918060 | 2022-04-20 22:06:55.126216 | 2022-04-20 22:07:25.126216 | 1 | {} 755046861432094721 | AUTO CREATE STATS | Table statistics refresh for movr.public.user_promo_codes | CREATE STATISTICS __auto__ FROM [111] WITH OPTIONS THROTTLING 0.9 AS OF SYSTEM TIME '-30s' | root | succeeded | NULL | 2022-04-20 22:06:55.817903 | 2022-04-20 22:06:55.825208 | 2022-04-20 22:06:55.858624 | 2022-04-20 22:06:55.857442 | 1 | | 1 | 944378671132247270 | 2022-04-20 22:06:55.825209 | 2022-04-20 22:07:25.825209 | 1 | {} 755046861588529153 | AUTO CREATE STATS | Table statistics refresh for movr.public.rides | CREATE STATISTICS __auto__ FROM [108] WITH OPTIONS THROTTLING 0.9 AS OF SYSTEM TIME '-30s' | root | succeeded | NULL | 2022-04-20 22:06:55.865648 | 2022-04-20 22:06:55.871917 | 2022-04-20 22:06:56.772123 | 2022-04-20 22:06:56.770818 | 1 | | 1 | 107998367520189635 | 2022-04-20 22:06:55.871917 | 2022-04-20 22:07:25.871917 | 1 | {} 755046864579690497 | AUTO CREATE STATS | Table statistics refresh for movr.public.vehicles | CREATE STATISTICS __auto__ FROM [107] WITH OPTIONS THROTTLING 0.9 AS OF SYSTEM TIME '-30s' | root | succeeded | NULL | 2022-04-20 22:06:56.778476 | 2022-04-20 22:06:56.785127 | 2022-04-20 22:06:56.887308 | 2022-04-20 22:06:56.886099 | 1 | | 1 | 1370771572055208171 | 2022-04-20 22:06:56.785128 | 2022-04-20 22:07:26.785128 | 1 | {} 755046864953376769 | AUTO CREATE STATS | Table statistics refresh for movr.public.promo_codes | CREATE STATISTICS __auto__ FROM [110] WITH OPTIONS THROTTLING 0.9 AS OF SYSTEM TIME '-30s' | root | succeeded | NULL | 2022-04-20 22:06:56.892515 | 2022-04-20 22:06:56.898684 | 2022-04-20 22:06:57.416186 | 2022-04-20 22:06:57.414741 | 1 | | 1 | 3756312875539405235 | 2022-04-20 22:06:56.898685 | 2022-04-20 22:07:26.898685 | 1 | {} 755046866689851393 | AUTO CREATE STATS | Table statistics refresh for movr.public.users | CREATE STATISTICS __auto__ FROM [106] WITH OPTIONS THROTTLING 0.9 AS OF SYSTEM TIME '-30s' | root | succeeded | NULL | 2022-04-20 22:06:57.422449 | 2022-04-20 22:06:57.428999 | 2022-04-20 22:06:57.573797 | 2022-04-20 22:06:57.57256 | 1 | | 1 | 6690084610338235566 | 2022-04-20 22:06:57.428999 | 2022-04-20 22:07:27.428999 | 1 | {} (6 rows)
Known limitations
The ANALYZE alias does not support specifying an AS OF SYSTEM TIME timestamp. ANALYZE statements use AS OF SYSTEM TIME '-0.001ms' automatically. For more control over the statistics interval, use the CREATE STATISTICS syntax instead. #96430
See also
Was this helpful?
