Version: Latest-3.5

Gather statistics for CBO

This topic describes the basic concept of StarRocks cost-based optimizer (CBO) and how to collect statistics for the CBO to select an optimal query plan. StarRocks 2.4 introduces histograms to gather accurate data distribution statistics.

Since v3.2.0, StarRocks supports collecting statistics from Hive, Iceberg, and Hudi tables, reducing the dependency on other metastore systems. The syntax is similar to collecting StarRocks internal tables.

What is CBO

CBO is critical to query optimization. After an SQL query arrives at StarRocks, it is parsed into a logical execution plan. The CBO rewrites and transforms the logical plan into multiple physical execution plans. The CBO then estimates the execution cost of each operator in the plan (such as CPU, memory, network, and I/O) and chooses a query path with the lowest cost as the final physical plan.

StarRocks CBO was launched in StarRocks 1.16.0 and is enabled by default from 1.19 onwards. Developed upon the Cascades framework, StarRocks CBO estimates cost based on various statistical information. It is capable of choosing an execution plan with the lowest cost among tens of thousands of execution plans, significantly improving the efficiency and performance of complex queries.

Statistics are important for the CBO. They determine whether cost estimation is accurate and useful. The following sections detail the types of statistical information, the collection policy, and how to collect statistics and view statistical information.

Types of statistical information

StarRocks collects a variety of statistics as the input for cost estimation.

Basic statistics

By default, StarRocks periodically collects the following basic statistics of tables and columns:

row_count: total number of rows in a table
data_size: data size of a column
ndv: column cardinality, which is the number of distinct values in a column
null_count: the amount of data with NULL values in a column
min: minimum value in a column
max: maximum value in a column

Full statistics are stored in the column_statistics of the _statistics_ database. You can view this table to query table statistics. Following is an example of querying statistics data from this table.

SELECT * FROM _statistics_.column_statistics\G
*************************** 1. row ***************************
      table_id: 10174
  partition_id: 10170
   column_name: is_minor
         db_id: 10169
    table_name: zj_test.source_wiki_edit
partition_name: p06
     row_count: 2
     data_size: 2
           ndv: NULL
    null_count: 0
           max: 1
           min: 0

Histogram

StarRocks 2.4 introduces histograms to complement basic statistics. Histogram is considered an effective way of data representation. For tables with skewed data, histograms can accurately reflect data distribution.

StarRocks uses equi-height histograms, which are constructed on several buckets. Each bucket contains an equal amount of data. For data values that are frequently queried and have a major impact on selectivity, StarRocks allocates separate buckets for them. More buckets mean more accurate estimation, but may also cause a slight increase in memory usage. You can adjust the number of buckets and most common values (MCVs) for a histogram collection task.

Histograms are applicable to columns with highly skewed data and frequent queries.If your table data is uniformly distributed, you do not need to create histograms. Histograms can be created only on columns of numeric, DATE, DATETIME, or string types.

Histograms are stored in the histogram_statistics table of the _statistics_ database. Following is an example of querying statistics data from this table:

SELECT * FROM _statistics_.histogram_statistics\G
*************************** 1. row ***************************
   table_id: 10174
column_name: added
      db_id: 10169
 table_name: zj_test.source_wiki_edit
    buckets: NULL
        mcv: [["23","1"],["5","1"]]
update_time: 2023-12-01 15:17:10.274000

Multi-column joint statistics

Since v3.5.0, StarRocks supports multi-column joint statistics collection. Currently, when StarRocks performs cardinality estimation, the optimizer in most scenarios assumes that multiple columns are completely independent of each other, that is, there is no correlation among them. However, if there is a correlation among columns, the current estimation method may lead to an incorrect result. This causes the optimizer to generate an incorrect execution plan. Currently, only multi-column joint NDV is supported, and it is mainly used in the following scenarios for cardinality estimation.

Evaluating multiple AND-joined equivalent predicates.
Evaluating Agg nodes.
Apply to the aggregate push-down strategy.

Currently, multi-column joint statistics collection only supports manual collection. The default type is sampled collection. Multi-column statistics are stored in the multi_column_statistics table of the _statistics_ database in each StarRocks cluster. A query will return information similar to the following:

mysql> select * from _statistics_.multi_column_statistics \G
*************************** 1. row ***************************
    table_id: 1695021
  column_ids: 0#1
       db_id: 110099
  table_name: db.test_multi_col_stats
column_names: id,name
         ndv: 11
 update_time: 2025-04-11 15:09:50

Collection types and methods

Data size and data distribution constantly change in a table. Statistics must be updated regularly to represent that data change. Before creating a statistics collection task, you must choose a collection type and method that best suit your business requirements.

StarRocks supports full and sampled collection, both can be performed automatically and manually. By default, StarRocks automatically collects full statistics of a table. It checks for any data updates every 5 minutes. If data change is detected, data collection will be automatically triggered. If you do not want to use automatic full collection, you can customize a collection task.

Collection type	Collection method	Description	Advantage and disadvantage
Full collection	Automatic/manual	Scans the full table to gather statistics.Statistics are collected partition. If a partition has no data change, data will not be collected from this partition, reducing resource consumption.Full statistics are stored in the `_statistics_.column_statistics` table.	Advantage: The statistics are accurate, which helps the CBO make accurate estimation.Disadvantage: It consumes system resources and is slow. From 2.5 onwards, StarRocks allows you to specify an automatic collection period, which reduces resource consumption.
Sampled collection	Automatic/manual	Evenly extracts `N` rows of data from each partition of a table.Statistics are collected by table. Basic statistics of each column are stored as one record. Cardinality information (ndv) of a column is estimated based on the sampled data, which is not accurate. Sampled statistics are stored in the `_statistics_.table_statistic_v1` table.	Advantage: It consumes less system resources and is fast.Disadvantage: The statistics are not complete, which may affect the accuracy of cost estimation.

Since v3.5.0, both sampled and full collection statistics are stored in the _statistics_.column_statistics table. This is because the current optimizer will focus on the most recently collected statistics in the cardinality estimation, while each time the background automatically collects statistics, it may use a different collection method due to the different health status of the table. If there is data skew, the error rate of the sampled statistical information of the full table may be higher, and the same query may use different statistical information due to the different collection methods, which may cause the optimizer to generate an incorrect execution plan. Therefore, both sampled collection and full collection assume that the statistical information is collected at the partition level. You can adjust to the previous collection and storage method by modifying the FE configuration item statistic_use_meta_statistics to false.

Predicate Column

Since v3.5.0, StarRocks supports collecting Predicate Column statistics.

A Predicate Column is a column that is often used as a filter condition (WHERE clause, JOIN condition, GROUP BY column, DISTINCT column) in a query. StarRocks automatically records each Predicate Column for the table involved in a query and stores it in _statistics_. predicate_columns table. The query will return the following:

select * from _statistics_.predicate_columns \G
*************************** 1. row ***************************
      fe_id: 127.0.0.1_9011_1735874871718
      db_id: 1684773
   table_id: 1685786
  column_id: 1
      usage: normal,predicate,join,group_by
  last_used: 2025-04-11 20:39:32
    created: 2025-04-11 20:39:53

You can also query the information_schema.column_stats_usage view for more observable information. In extreme scenarios (with a large number of columns), there is a large overhead for full collection, whereas in reality, for a stable workload, you often do not need statistics for all columns, but only the columns involved in some key Filter, Join, and Aggregation. Therefore, in order to weigh the cost and accuracy, StarRocks supports collecting Predicate Column statistics manually, and collecting Predicate Column statistics according to the policy when collecting statistics automatically, avoiding collecting all column statistics in the table. Each FE node in the cluster synchronously updates the Predicate Column information to the FE cache at regular intervals to accelerate the usage.

Collect statistics

StarRocks offers flexible statistics collection methods. You can choose automatic, manual, or custom collection, whichever suits your business scenarios.

By default, StarRocks automatically collects the full amount of statistics for a table on a periodic basis. The default interval is 10 minutes. If the system finds that the update ratio of the data meets the condition, the collection will be triggered automatically. Collecting full statistics may consume a lot of system resources. If you do not want to use automatic full statistics collection, you can set the FE configuration item enable_collect_full_statistic to false, and the periodic collection task will be changed from full statistics collection to sampled collection.

Since v3.5.0, StarRocks will automatically change the full collection to sampled collection if it finds that the data of a table has changed a lot since the last collection to the current table during automatic collection. In the case of sampled collection is used, if there is a Predicate Column (PREDICATE/JOIN/GROUP BY/DISTINCT) in the table, StarRocks will convert the sampling task to full collection and collect the statistics of the Predicate Column to ensure the accuracy of the statistics, and will not collect the statistics of all columns in the table. It will not collect statistics from all columns in the table. You can configure this with the FE configuration item statistic_auto_collect_use_full_predicate_column_for_sample. In addition, StarRocks also switches from full collection of all columns to full collection of Predicate Columns if the number of columns in the table exceeds the FE configuration item statistic_auto_collect_predicate_columns_threshold when full collection is used.

Automatic collection

For basic statistics, StarRocks automatically collects full statistics of a table by default, without requiring manual operations. For tables on which no statistics have been collected, StarRocks automatically collects statistics within the scheduling period. For tables on which statistics have been collected, StarRocks updates the total number of rows and modified rows in the tables, and persists this information regularly for judging whether to trigger automatic collection. Histograms and multi-column joint statistics are not captured by the current automatic collection task.

From 2.4.5 onwards, StarRocks allows you to specify a collection period for automatic full collection, which prevents cluster performance jitter caused by automatic full collection. This period is specified by FE parameters statistic_auto_analyze_start_time and statistic_auto_analyze_end_time.

Conditions that will trigger automatic collection:

Table data has changed since previous statistics collection.
The collection time falls within the range of the configured collection period. (The default collection period is all day.)
The update time of the previous collecting job is earlier than the latest update time of partitions.
The health of table statistics is below the specified threshold (statistic_auto_collect_ratio).

Formula for calculating statistics health:

If the number of partitions with data updated is less than 10, the formula is 1 - (Number of updated rows since previous collection/Total number of rows).

If the number of partitions with data updated is greater than or equal to 10, the formula is 1 - MIN(Number of updated rows since previous collection/Total number of rows, Number of updated partitions since previous collection/Total number of partitions).

Since v3.5.0, in order to determine whether a partition is healthy or not, StarRocks no longer compares the statistic information collection time with the data update time, but rather the ratio of partition update rows. You can configure this via the FE configuration item statistic_partition_health_v2_threshold. Also, you can set the FE configuration item statistic_partition_healthy_v2 to false to use to the previous health check behavior.

In addition, StarRocks allows you to configure collection policies based on table size and table update frequency:

For tables with small data volume, **statistics are collected in real time with no restrictions, even though the table data is frequently updated. The statistic_auto_collect_small_table_size parameter can be used to determine whether a table is a small table or a large table. You can also use statistic_auto_collect_small_table_interval to configure collection intervals for small tables.
For tables with large data volume, the following restrictions apply:
- The default collection interval is not less than 12 hours, which can be configured using statistic_auto_collect_large_table_interval.
- When the collection interval is met and the statistics health is lower than the threshold for automatic sampled collection, sampled collection is triggered. You can configure this behavior using the FE configurartion item statistic_auto_collect_sample_threshold. Since v3.5.0, sampled collection is converted to full collection of Predicate Columns if all of the following conditions are met:
  - Predicate Column exists in the table, and the number of Predicate Columns is less than statistic_auto_collect_max_predicate_column_size_on_sample_strategy.
  - The FE configuration statistic_auto_collect_use_full_predicate_column_for_sample is set to true.
- When the collection interval is met and the statistics health is higher than the threshold for automatic sampled collection (statistic_auto_collect_sample_threshold) and lower than the automatic collection threshold (statistic_auto_collect_ratio), full collection is triggered.
- When the size of partitions to collect data (statistic_max_full_collect_data_size) is greater than 100 GB, sampled collection is triggered.
- Only statistics of partitions whose update time is later than the time of the previous collection task are collected. Statistics of partitions with no data change are not collected.
For tables that have a Predicate Column and the total number of columns exceeds the statistic_auto_collect_predicate_columns_threshold, statistics are only collected for the Predicate Column in the table.

tip

After the data of a table is changed, manually triggering a sampled collection task for this table will make the update time of the sampled collection task later than the data update time, which will not trigger automatic full collection for this table in this scheduling period.

Automatic full collection is enabled by default and run by the system using the default settings.

The following table describes the default settings. If you need to modify them, run the ADMIN SET CONFIG command.

FE configuration item	Type	Default value	Description
enable_statistic_collect	BOOLEAN	TRUE	Whether to turn on the default and user-defined auto collection tasks. This switch is turned on by default.
enable_collect_full_statistic	BOOLEAN	TRUE	Whether to enable the default auto collection. This switch is turned on by default.
statistic_collect_interval_sec	LONG	600	The interval for checking data updates during automatic collection. Unit: seconds.
statistic_auto_analyze_start_time	STRING	00:00:00	The start time of automatic collection. Value range: `00:00:00` - `23:59:59`.
statistic_auto_analyze_end_time	STRING	23:59:59	The end time of automatic collection. Value range: `00:00:00` - `23:59:59`.
statistic_auto_collect_small_table_size	LONG	5368709120	The threshold for determining whether a table is a small table for automatic full collection. A table whose size is greater than this value is considered a large table, whereas a table whose size is less than or equal to this value is considered a small table. Unit: Byte. Default value: 5368709120 (5 GB).
statistic_auto_collect_small_table_interval	LONG	0	The interval for automatically collecting full statistics of small tables. Unit: seconds.
statistic_auto_collect_large_table_interval	LONG	43200	The interval for automatically collecting full statistics of large tables. Unit: seconds. Default value: 43200 (12 hours).
statistic_auto_collect_ratio	FLOAT	0.8	The threshold for determining whether the statistics for automatic collection are healthy. If statistics health is below this threshold, automatic collection is triggered.
statistic_auto_collect_sample_threshold	DOUBLE	0.3	The statistics health threshold for triggering automatic sampled collection. If the health value of statistics is lower than this threshold, automatic sampled collection is triggered.
statistic_max_full_collect_data_size	LONG	107374182400	The data size of the partitions for automatic collection to collect data. Unit: Byte. Default value: 107374182400 (100 GB). If the data size exceeds this value, full collection is discarded and sampled collection is performed instead.
statistic_full_collect_buffer	LONG	20971520	The maximum buffer size taken by automatic collection tasks. Unit: Byte. Default value: 20971520 (20 MB).
statistic_collect_max_row_count_per_query	INT	5000000000	The maximum number of rows to query for a single analyze task. An analyze task will be split into multiple queries if this value is exceeded.
statistic_collect_too_many_version_sleep	LONG	600000	The sleep time of automatic collection tasks if the table on which the collection task runs has too many data versions. Unit: ms. Default value: 600000 (10 minutes).
statistic_auto_collect_use_full_predicate_column_for_sample	BOOLEAN	TRUE	Whether to convert an automatic full collection task to full collection of a Predicate Column when it hits the sampled collection policy.
statistic_auto_collect_max_predicate_column_size_on_sample_strategy	INT	16	When the Auto Full Collection task hits the sampled collection policy, if the table has an unusually large number of Predicate Columns and exceeds this configuration item, the task will not switch to full collection of the Predicate Columns, but maintains to be the sampled collection of all columns. This configuration item controls the maximum value of Predicate Column for this behavior.
statistic_auto_collect_predicate_columns_threshold	INT	32	If the number of columns in the table exceeds this configuration during automatic collection, only the column statistics for the Predicate Column will be collected.
statistic_predicate_columns_persist_interval_sec	LONG	60	The interval at which FE synchronize and persists statistics of Predicate Column.
statistic_predicate_columns_ttl_hours	LONG	24	The elimination time of the Predicate Column statistics cached in FE.

You can rely on automatic jobs for a majority of statistics collection, but if you have specific requirements, you can manually create a task by executing the ANALYZE TABLE statement or customize an automatic task by executing the CREATE ANALYZE statement.

Manual collection

You can use ANALYZE TABLE to create a manual collection task. By default, manual collection is a synchronous operation. You can also set it to an asynchronous operation. In asynchronous mode, after you run ANALYZE TABLE, the system immediately returns whether this statement is successful. However, the collection task will be running in the background and you do not have to wait for the result. You can check the status of the task by running SHOW ANALYZE STATUS. Asynchronous collection is suitable for tables with large data volume, whereas synchronous collection is suitable for tables with small data volume. Manual collection tasks are run only once after creation. You do not need to delete manual collection tasks. You must have the INSERT and SELECT privileges on the coreesponding table to perform the ANALYZE TABLE operation.

Manually collect basic statistics

ANALYZE [FULL|SAMPLE] TABLE tbl_name 
    [( col_name [, col_name]... )
    | col_name [, col_name]...
    | ALL COLUMNS
    | PREDICATE COLUMNS
    | MULTIPLE COLUMNS ( col_name [, col_name]... )]
[PARTITION (partition_name [, partition_name]...)]
[WITH [SYNC | ASYNC] MODE]
[PROPERTIES (property [, property]...)]

Parameter description:

Collection type
- FULL: indicates full collection.
- SAMPLE: indicates sampled collection.
- If no collection type is specified, full collection is used by default.
The type of the column from which to collect statistics:
- col_name: Columns from which to collect statistics. Separate multiple columns with commas (,). If this parameter is not specified, the entire table is collected.
- ALL COLUMNS: Collect statistics from all columns. Supported since v3.5.0.
- PREDICATE COLUMNS: Collect statistics from only Predicate Columns. Supported since v3.5.0.
- MULTIPLE COLUMNS: Collects joint statistics from the specified multiple columns. Currently, only manual synchronous collection of multiple columns is supported. The number of columns for manual statistics collection cannot exceed statistics_max_multi_column_combined_num, the default value is 10. Supported since v3.5.0.
[WITH SYNC | ASYNC MODE]: whether to run the manual collection task in synchronous or asynchronous mode. Synchronous collection is used by default if you do not specify this parameter.
PROPERTIES: custom parameters. If PROPERTIES is not specified, the default settings in the fe.conf file are used. The properties that are actually used can be viewed via the Properties column in the output of SHOW ANALYZE STATUS.

PROPERTIES	Type	Default value	Description
statistic_sample_collect_rows	INT	200000	The minimum number of rows to collect for sampled collection.If the parameter value exceeds the actual number of rows in your table, full collection is performed.

Examples

Manual full collection

-- Manually collect full stats of a table using default settings.
ANALYZE TABLE tbl_name;

-- Manually collect full stats of a table using default settings.
ANALYZE FULL TABLE tbl_name;

-- Manually collect stats of specified columns in a table using default settings.
ANALYZE TABLE tbl_name(c1, c2, c3);

Manual sampled collection

-- Manually collect partial stats of a table using default settings.
ANALYZE SAMPLE TABLE tbl_name;

-- Manually collect stats of specified columns in a table, with the number of rows to collect specified.
ANALYZE SAMPLE TABLE tbl_name (v1, v2, v3) PROPERTIES(
    "statistic_sample_collect_rows" = "1000000"
);

Manual collection of multi-column joint statistics

-- Manual sampled collection of multi-column joint statistics
ANALYZE SAMPLE TABLE tbl_name MULTIPLE COLUMNS (v1, v2);

-- Manual full collection of multi-column joint statistics
ANALYZE FULL TABLE tbl_name MULTIPLE COLUMNS (v1, v2);

Manual collection of Predicate Column

-- Manual sampled collection of Predicate Column
ANALYZE SAMPLE TABLE tbl_name PREDICATE COLUMNS

-- Manual full collection of Predicate Column
ANALYZE FULL TABLE tbl_name PREDICATE COLUMNS

Manually collect histograms

ANALYZE TABLE tbl_name UPDATE HISTOGRAM ON col_name [, col_name]
[WITH SYNC | ASYNC MODE]
[WITH N BUCKETS]
[PROPERTIES (property [,property])]

Parameter description:

col_name: columns from which to collect statistics. Separate multiple columns with commas (,). If this parameter is not specified, the entire table is collected. This parameter is required for histograms.
[WITH SYNC | ASYNC MODE]: whether to run the manual collection task in synchronous or asynchronous mode. Synchronous collection is used by default if you not specify this parameter.
WITH N BUCKETS: N is the number of buckets for histogram collection. If not specified, the default value in fe.conf is used.
PROPERTIES: custom parameters. If PROPERTIES is not specified, the default settings in fe.conf are used.

PROPERTIES	Type	Default value	Description
statistic_sample_collect_rows	INT	200000	The minimum number of rows to collect. If the parameter value exceeds the actual number of rows in your table, full collection is performed.
histogram_buckets_size	LONG	64	The default bucket number for a histogram.
histogram_mcv_size	INT	100	The number of most common values (MCV) for a histogram.
histogram_sample_ratio	FLOAT	0.1	The sampling ratio for a histogram.
histogram_max_sample_row_count	LONG	10000000	The maximum number of rows to collect for a histogram.

The number of rows to collect for a histogram is controlled by multiple parameters. It is the larger value between statistic_sample_collect_rows and table row count * histogram_sample_ratio. The number cannot exceed the value specified by histogram_max_sample_row_count. If the value is exceeded, histogram_max_sample_row_count takes precedence.

The properties that are actually used can be viewed via the Properties column in the output of SHOW ANALYZE STATUS.

Examples

-- Manually collect histograms on v1 using the default settings.
ANALYZE TABLE tbl_name UPDATE HISTOGRAM ON v1;

-- Manually collect histograms on v1 and v2, with 32 buckets, 32 MCVs, and 50% sampling ratio.
ANALYZE TABLE tbl_name UPDATE HISTOGRAM ON v1,v2 WITH 32 BUCKETS 
PROPERTIES(
   "histogram_mcv_size" = "32",
   "histogram_sample_ratio" = "0.5"
);

Custom collection

Customize an automatic collection task

The default collection tasks provided by StarRocks automatically collect statistics for all databases and all tables according to the policy, and by default, you do not need to create custom collection tasks.

If you want to customize an automatic collection task, you need to create it through the CREATE ANALYZE statement. To create a collection task, you need the INSERT and SELECT permissions for the table being collected.

-- Automatically collect stats of all databases.
CREATE ANALYZE [FULL|SAMPLE] ALL [PROPERTIES (property [,property])]

-- Automatically collect stats of all tables in a database.
CREATE ANALYZE [FULL|SAMPLE] DATABASE db_name
[PROPERTIES (property [,property])]

-- Automatically collect stats of specified columns in a table.
CREATE ANALYZE [FULL|SAMPLE] TABLE tbl_name (col_name [,col_name])
[PROPERTIES (property [,property])]

-- Automatically collect histograms of specified columns in a table.
CREATE ANALYZE TABLE tbl_name UPDATE HISTOGRAM ON col_name [, col_name]
[WITH SYNC | ASYNC MODE]
[WITH N BUCKETS]
[PROPERTIES (property [,property])]

Parameter description:

Collection type
- FULL: indicates full collection.
- SAMPLE: indicates sampled collection.
- If no collection type is specified, sample collection is used by default.
col_name: columns from which to collect statistics. Separate multiple columns with commas (,). If this parameter is not specified, the entire table is collected.
PROPERTIES: custom parameters. If PROPERTIES is not specified, the default settings in fe.conf are used.

PROPERTIES	Type	Default value	Description
statistic_auto_collect_ratio	FLOAT	0.8	The threshold for determining whether the statistics for automatic collection are healthy. If the statistics health is below this threshold, automatic collection is triggered.
statistic_sample_collect_rows	INT	200000	The minimum number of rows to collect.If the parameter value exceeds the actual number of rows in your table, full collection is performed.
statistic_exclude_pattern	String	null	The name of the database or table that needs to be excluded in the job. You can specify the database and table that do not collect statistics in the job. Note that this is a regular expression pattern, and the match content is `database.table`.
statistic_auto_collect_interval	LONG	0	The interval for automatic collection. Unit: seconds. By default, StarRocks chooses `statistic_auto_collect_small_table_interval` or `statistic_auto_collect_large_table_interval` as the collection interval based on the table size. If you specified the `statistic_auto_collect_interval` property when you create an analyze job, this setting takes precedence over `statistic_auto_collect_small_table_interval` and `statistic_auto_collect_large_table_interval`.

Examples

Automatic full collection

-- Automatically collect full stats of all databases.
CREATE ANALYZE ALL;

-- Automatically collect full stats of a database.
CREATE ANALYZE DATABASE db_name;

-- Automatically collect full stats of all tables in a database.
CREATE ANALYZE FULL DATABASE db_name;

-- Automatically collect full stats of specified columns in a table.
CREATE ANALYZE TABLE tbl_name(c1, c2, c3); 

-- Automatically collect stats of all databases, excluding specified database 'db_name'.
CREATE ANALYZE ALL PROPERTIES (
   "statistic_exclude_pattern" = "db_name\."
);

-- Automatically collect histograms of specified database, table, or columns.
CREATE ANALYZE TABLE tbl_name UPDATE HISTOGRAM ON c1,c2;

Automatic sampled collection

-- Automatically collect stats of all tables in a database with default settings.
CREATE ANALYZE SAMPLE DATABASE db_name;

-- Automatically collect stats of all tables in a database, excluding specified table 'db_name.tbl_name'.
CREATE ANALYZE SAMPLE DATABASE db_name PROPERTIES (
   "statistic_exclude_pattern" = "db_name.tbl_name"
);

-- Automatically collect stats of specified columns in a table, with statistics health and the number of rows to collect specified.
CREATE ANALYZE SAMPLE TABLE tbl_name(c1, c2, c3) PROPERTIES (
   "statistic_auto_collect_ratio" = "0.5",
   "statistic_sample_collect_rows" = "1000000"
);

Instead of the auto collection task provided by StarRocks, use a user-defined collection task that does not collect the db_name.tbl_name table.

ADMIN SET FRONTEND CONFIG("enable_auto_collect_statistics"="false");
DROP ALL ANALYZE JOB;
CREATE ANALYZE FULL ALL db_name PROPERTIES (
   "statistic_exclude_pattern" = "db_name.tbl_name"
);

Collect statistics during data loading

To ensure a good execution plan for queries immediately after data loading, StarRocks triggers an asynchronous statistic collection task at the end of the INSERT INTO/OVERWRITE DML statements, which by default waits for 30 seconds after the DML finishes. If the statistics collection task does not finish in 30 seconds, the DML execution result is returned.

INSERT INTO

Statistics is collected only for the first import of data into the partition.
If the number of rows in this import is greater than statistic_sample_collect_rows, the sampled collection task is triggered, otherwise full collection is used.

INSERT OVERWRITE

If the ratio of the change in the number of rows before and after the OVERWRITE is less than statistic_sample_collect_ratio_threshold_of_first_load, the statistics collection task is not triggered.
If the number of rows in this OVERWRITE operation is greater than statistic_sample_collect_rows, then a sampled collection task is triggered, otherwise full collection is used.

The following properties (PROPERTIES) are used to create customized collection tasks for data loading. If not configured, the value of the corresponding FE configuration item is used.

PROPERTIES	Type	Dedault	Description
enable_statistic_collect_on_first_load	BOOLEAN	TRUE	Whether to trigger the statistics collection task after executing INSERT INTO/OVERWRITE.
semi_sync_collect_statistic_await_seconds	LONG	30	Maximum time to wait for statistics to be collected before returning results.
statistic_sample_collect_ratio_threshold_of_first_load	DOUBLE	0.1	The ratio of data changes in OVERWRITE operations that will not trigger statistics collection tasks.
statistic_sample_collect_rows	LONG	200000	When the total data rows loaded via DML statements exceeds this value, sampled collection is used for statistics collection.

View custom collection tasks

SHOW ANALYZE JOB [WHERE predicate][ORDER BY columns][LIMIT num]

You can filter results by using the WHERE clause. The statement returns the following columns.

Column	Description
Id	The ID of the collection task.
Database	The database name.
Table	The table name.
Columns	The column names.
Type	The type of statistics, including `FULL` and `SAMPLE`.
Schedule	The type of scheduling. The type is `SCHEDULE` for an automatic task.
Properties	Custom parameters.
Status	The task status, including PENDING, RUNNING, SUCCESS, and FAILED.
LastWorkTime	The time of the last collection.
Reason	The reason why the task failed. NULL is returned if task execution was successful.

Examples

-- View all the custom collection tasks.
SHOW ANALYZE JOB

-- View custom collection tasks of database `test`.
SHOW ANALYZE JOB where `database` = 'test';

Delete a custom collection task

DROP ANALYZE <ID>
| DROP ALL ANALYZE JOB

The task ID can be obtained by using the SHOW ANALYZE JOB statement.

Examples

DROP ANALYZE 266030;

DROP ALL ANALYZE JOB;

View status of collection tasks

You can view the status of all the current tasks by running the SHOW ANALYZE STATUS statement. This statement cannot be used to view the status of custom collection tasks. To view the status of custom collection tasks, use SHOW ANALYZE JOB.

SHOW ANALYZE STATUS [WHERE predicate];

You can use LIKE or WHERE to filter the information to return.

This statement returns the following columns.

List name	Description
Id	The ID of the collection task.
Database	The database name.
Table	The table name.
Columns	The column names.
Type	The type of statistics, including FULL, SAMPLE, and HISTOGRAM.
Schedule	The type of scheduling. `ONCE` means manual and `SCHEDULE` means automatic.
Status	The status of the task.
StartTime	The time when the task starts executing.
EndTime	The time when the task execution ends.
Properties	Custom parameters.
Reason	The reason why the task failed. NULL is returned if the execution was successful.

View statistics

View metadata of basic statistics

SHOW STATS META [WHERE predicate][ORDER BY columns][LIMIT num]

This statement returns the following columns.

Column	Description
Database	The database name.
Table	The table name.
Columns	The column names.
Type	The type of statistics. `FULL` means full collection and `SAMPLE` means sampled collection.
UpdateTime	The latest statistics update time for the current table.
Properties	Custom parameters.
Healthy	The health of statistical information.
ColumnStats	Column ANALYZE type.
TabletStatsReportTime	The time when the Tablet metadata for the table was updated in the FE.
TableHealthyMetrics	Metric of health in statistical information.
TableUpdateTime	The time when the table was updated.

View metadata of histograms

SHOW HISTOGRAM META [WHERE predicate]

This statement returns the following columns.

Column	Description
Database	The database name.
Table	The table name.
Column	The columns.
Type	Type of statistics. The value is `HISTOGRAM` for histograms.
UpdateTime	The latest statistics update time for the current table.
Properties	Custom parameters.

Delete statistics

You can delete statistical information you do not need. When you delete statistics, both the data and metadata of the statistics are deleted, as well as the statistics in expired cache. Note that if an automatic collection task is ongoing, previously deleted statistics may be collected again. You can use SHOW ANALYZE STATUS to view the history of collection tasks.

Delete basic statistics

The following statement deletes the statistics stored in the default_catalog._statistics_.column_statistics table, and the corresponding table statistics cached by FE will also be invalidated. Since v3.5.0, this statement also deletes multi-column joint statistics for this table.

DROP STATS tbl_name

The following statement deletes the multi-column joint statistics stored in the default_catalog._statistics_.multi_column_statistics table, and the multi-column joint statistics in the corresponding table cached by FE will also be invalidated. This statement does not delete the basic statistics of the table.

DROP MULTIPLE COLUMNS STATS tbl_name

Delete histograms

ANALYZE TABLE tbl_name DROP HISTOGRAM ON col_name [, col_name]

Cancel a collection task

You can use the KILL ANALYZE statement to cancel a running collection task, including manual and custom tasks.

KILL ANALYZE <ID>

The task ID for a manual collection task can be obtained from SHOW ANALYZE STATUS. The task ID for a custom collection task can be obtained from SHOW ANALYZE SHOW ANALYZE JOB.

Other FE configuration items

FE configuration item	Type	Default value	Description
statistic_manager_sleep_time_sec	LONG	60	The interval at which metadata is scheduled. Unit: seconds. The system performs the following operations based on this interval:Create tables for storing statistics.Delete statistics that have been deleted.Delete expired statistics.
statistic_analyze_status_keep_second	LONG	259200	The duration to retain the history of collection tasks. Unit: seconds. Default value: 259200 (3 days).

Session variable

statistic_collect_parallel: Used to adjust the parallelism of statistics collection tasks that can run on BEs. Default value: 1. You can increase this value to speed up collection tasks.

Collect statistics of external tables

Since v3.2.0, StarRocks supports collecting statistics of Hive, Iceberg, and Hudi tables. The syntax is similar to collecting StarRocks internal tables. However, only manual full collection, manual histogram collection (since v3.2.7), and automatic full collection are supported. Sampled collection is not supported. Since v3.3.0, StarRocks supports collecting statistics of Delta Lake tables and statistics of sub-fields in STRUCT. Since v3.4.0, StarRocks supports automatic statistics collection via query-triggered ANALYZE tasks.

The collected statistics are stored in the external_column_statistics table of the _statistics_ in the default_catalog. They are not stored in Hive Metastore and cannot be shared by other search engines. You can query data from the default_catalog._statistics_.external_column_statistics table to verify whether statistics are collected for a Hive/Iceberg/Hudi table.

Following is an example of querying statistics data from external_column_statistics.

SELECT * FROM _statistics_.external_column_statistics\G
*************************** 1. row ***************************
    table_uuid: hive_catalog.tn_test.ex_hive_tbl.1673596430
partition_name: 
   column_name: k1
  catalog_name: hive_catalog
       db_name: tn_test
    table_name: ex_hive_tbl
     row_count: 3
     data_size: 12
           ndv: NULL
    null_count: 0
           max: 3
           min: 2
   update_time: 2023-12-01 14:57:27.137000

Limits

The following limits apply when you collect statistics for external tables:

You can collect statistics of only Hive, Iceberg, Hudi, and Delta Lake (Since v3.3.0) tables.
Only manual full collection, manual histogram collection (since v3.2.7), and automatic full collection are supported. Sampled collection is not supported.
For the system to automatically collect full statistics, you must create an Analyze job, which is different from collecting statistics of StarRocks internal tables where the system does this in the background by default.
For automatic collection tasks:
- You can only collect statistics of a specific table. You cannot collect statistics of all tables in a database or statistics of all databases in an external catalog.
- StarRocks can detect whether data in Hive and Iceberg tables are updated and if so, collect statistics of only partitions whose data is updated. StarRocks cannot perceive whether data in Hudi tables are updated and can only perform periodic full collection.
For query-triggered collection tasks:
- Currently, only Leader FE node can trigger ANALYZE tasks.
- The system only supports checking for partition changes on Hive and Iceberg tables, and only collects statistics on the partitions where the data has changed. For the Delta Lake/Hudi tables, the system collects statistics of the entire table.
- If Partition Transforms are applied to Iceberg tables, the collection of statistics is supported only for identity, year, month, day, hour type Transforms.
- Collecting statistics for Partition Evolution for Iceberg tables is not supported.

The following examples happen in a database under the Hive external catalog. If you want to collect statistics of a Hive table from the default_catalog, reference the table in the [catalog_name.][database_name.]<table_name> format.

Query-triggered collection

Since v3.4.0, the system supports automatic statistics collection of external tables by query-triggered ANALYZE tasks. When querying Hive, Iceberg, Hudi, or Delta Lake tables, the system will automatically trigger ANALYZE tasks in the background to collect statistics of the corresponding tables and columns, and use it for subsequent query plan optimization.

Workflow:

When the optimizer queries the cached statistics in FE, it will determine the object of the ANALYZE task based on the queried tables and columns (the ANALYZE task will only collect the statistics of the columns included in the query).
The system will encapsulate the task object as an ANALYZE task and add it to the PendingTaskQueue.
The Scheduler thread periodically fetches tasks from the PendingTaskQueue and add them to the RunningTasksQueue.
During the execution of the ANALYZE task, it collects statistics and writes them to the BE and clears the expired statistics cached in the FE.

This feature is enabled by default. You can control the above process with the following system variables and configuration items.

System variable

enable_query_trigger_analyze

Default: true
Type: Boolean
Description: Whether to enable query-trigger ANALYZE tasks.
Introduced in: v3.4.0

FE configuration

connector_table_query_trigger_analyze_small_table_rows

Default: 10000000
Type: Int
Unit: -
Is mutable: Yes
Description: The threshold for determining whether a table is a small table for query-trigger ANALYZE tasks.
Introduced in: v3.4.0

connector_table_query_trigger_analyze_small_table_interval

Default: 2 * 3600
Type: Int
Unit: Second
Is mutable: Yes
Description: The interval for query-trigger ANALYZE tasks of small tables.
Introduced in: v3.4.0

connector_table_query_trigger_analyze_large_table_interval

Default: 12 * 3600
Type: Int
Unit: Second
Is mutable: Yes
Description: The interval for query-trigger ANALYZE tasks of large tables.
Introduced in: v3.4.0

connector_table_query_trigger_analyze_max_pending_task_num

Default: 100
Type: Int
Unit: -
Is mutable: Yes
Description: Maximum number of query-trigger ANALYZE tasks that are in Pending state on the FE.
Introduced in: v3.4.0

connector_table_query_trigger_analyze_schedule_interval

Default: 30
Type: Int
Unit: Second
Is mutable: Yes
Description: The interval at which the Scheduler thread schedules to query-trigger ANALYZE tasks.
Introduced in: v3.4.0

connector_table_query_trigger_analyze_max_running_task_num

Default: 2
Type: Int
Unit: -
Is mutable: Yes
Description: Maximum number of query-trigger ANALYZE tasks that are in Running state on the FE.
Introduced in: v3.4.0

Manual collection

You can create an Analyze job on demand and the job runs immediately after you create it.

Create a manual collection task

Syntax:

-- Manual full collection
ANALYZE [FULL] TABLE tbl_name (col_name [,col_name])
[WITH SYNC | ASYNC MODE]
[PROPERTIES(property [,property])]

-- Manual histogram collection (since v3.3.0)
ANALYZE TABLE tbl_name UPDATE HISTOGRAM ON col_name [, col_name]
[WITH SYNC | ASYNC MODE]
[WITH N BUCKETS]
[PROPERTIES (property [,property])]

Following is an example of manual full collection:

ANALYZE TABLE ex_hive_tbl(k1);
+----------------------------------+---------+----------+----------+
| Table                            | Op      | Msg_type | Msg_text |
+----------------------------------+---------+----------+----------+
| hive_catalog.tn_test.ex_hive_tbl | analyze | status   | OK       |
+----------------------------------+---------+----------+----------+

View the task status

Syntax:

SHOW ANALYZE STATUS [LIKE | WHERE predicate]

Example:

SHOW ANALYZE STATUS where `table` = 'ex_hive_tbl';
+-------+----------------------+-------------+---------+------+----------+---------+---------------------+---------------------+------------+--------+
| Id    | Database             | Table       | Columns | Type | Schedule | Status  | StartTime           | EndTime             | Properties | Reason |
+-------+----------------------+-------------+---------+------+----------+---------+---------------------+---------------------+------------+--------+
| 16400 | hive_catalog.tn_test | ex_hive_tbl | k1      | FULL | ONCE     | SUCCESS | 2023-12-04 16:31:42 | 2023-12-04 16:31:42 | {}         |        |
| 16465 | hive_catalog.tn_test | ex_hive_tbl | k1      | FULL | ONCE     | SUCCESS | 2023-12-04 16:37:35 | 2023-12-04 16:37:35 | {}         |        |
| 16467 | hive_catalog.tn_test | ex_hive_tbl | k1      | FULL | ONCE     | SUCCESS | 2023-12-04 16:37:46 | 2023-12-04 16:37:46 | {}         |        |
+-------+----------------------+-------------+---------+------+----------+---------+---------------------+---------------------+------------+--------+

View metadata of statistics

Syntax:

SHOW STATS META [WHERE predicate]

Example:

SHOW STATS META where `table` = 'ex_hive_tbl';
+----------------------+-------------+---------+------+---------------------+------------+---------+
| Database             | Table       | Columns | Type | UpdateTime          | Properties | Healthy |
+----------------------+-------------+---------+------+---------------------+------------+---------+
| hive_catalog.tn_test | ex_hive_tbl | k1      | FULL | 2023-12-04 16:37:46 | {}         |         |
+----------------------+-------------+---------+------+---------------------+------------+---------+

Cancel a collection task

Cancels a running collection task.

Syntax:

KILL ANALYZE <ID>

You can view the task ID in the output of SHOW ANALYZE STATUS.

Automatic collection

For the system to automatically collect statistics of tables in an external data source, you can create an Analyze job. StarRocks automatically checks whether to run the task at the default check interval of 5 minutes. For Hive and Iceberg tables, StarRocks runs a collection task only when data in the tables are updated.

However, data changes in Hudi tables cannot be perceived and StarRocks periodically collects statistics based on the check interval and collection interval you specified. You can specify the following FE configuration items to control the collection behaviors:

statistic_collect_interval_sec

The interval for checking data updates during automatic collection. Unit: seconds. Default: 5 minutes.
statistic_auto_collect_small_table_rows (v3.2 and later)

Threshold to determine whether a table in an external data source (Hive, Iceberg, Hudi) is a small table during automatic collection. Default: 10000000.
statistic_auto_collect_small_table_interval

The interval for collecting statistics of small tables. Unit: seconds. Default: 0.
statistic_auto_collect_large_table_interval

The interval for collecting statistics of large tables. Unit: seconds. Default: 43200 (12 hours).

Automatic collection threads checks data updates at the interval specified by statistic_collect_interval_sec. If the number of rows in a table is less than statistic_auto_collect_small_table_rows, it collects statistics of such tables based on statistic_auto_collect_small_table_interval.

If the number of rows in a table exceeds statistic_auto_collect_small_table_rows, it collects statistics of such tables based on statistic_auto_collect_large_table_interval. It updates statistics of large tables only when Last table update time + Collection interval > Current time. This prevents frequent analyze tasks for large tables.

Create an automatic collection task

Syntax:

CREATE ANALYZE TABLE tbl_name (col_name [,col_name])
[PROPERTIES (property [,property])]

You can specify the property statistic_auto_collect_interval to set the collection interval exclusively for the automatic collection task. The FE configuration items statistic_auto_collect_small_table_interval and statistic_auto_collect_large_table_interval will not take effect on this task.

Example:

CREATE ANALYZE TABLE ex_hive_tbl (k1)
PROPERTIES ("statistic_auto_collect_interval" = "5");

Query OK, 0 rows affected (0.01 sec)

View the status of an automatic collection task

Same as manual collection.

View metadata of statistics

Same as manual collection.

View automatic collection tasks

Syntax:

SHOW ANALYZE JOB [WHERE predicate]

Example:

SHOW ANALYZE JOB WHERE `id` = '17204';

Empty set (0.00 sec)

Cancel a collection task

Same as manual collection.

Delete statistics

DROP STATS tbl_name

References

To query FE configuration items, run ADMIN SHOW CONFIG.
To modify FE configuration items, run ADMIN SET CONFIG.

What is CBO​

Types of statistical information​

Basic statistics​

Histogram​

Multi-column joint statistics​

Collection types and methods​

Predicate Column​

Collect statistics​

Automatic collection​

Manual collection​

Manually collect basic statistics​

Manually collect histograms​

Custom collection​

Customize an automatic collection task​

Collect statistics during data loading​

INSERT INTO​

INSERT OVERWRITE​

View custom collection tasks​

Delete a custom collection task​

View status of collection tasks​

View statistics​

View metadata of basic statistics​

View metadata of histograms​

Delete statistics​

Delete basic statistics​

Delete histograms​

Cancel a collection task​

Other FE configuration items​

Session variable​

Collect statistics of external tables​

Limits​

Query-triggered collection​

System variable​

enable_query_trigger_analyze​

FE configuration​

connector_table_query_trigger_analyze_small_table_rows​

connector_table_query_trigger_analyze_small_table_interval​

connector_table_query_trigger_analyze_large_table_interval​

connector_table_query_trigger_analyze_max_pending_task_num​

connector_table_query_trigger_analyze_schedule_interval​

connector_table_query_trigger_analyze_max_running_task_num​

Manual collection​

Create a manual collection task​

View the task status​

View metadata of statistics​

Cancel a collection task​

Automatic collection​

Create an automatic collection task​

View the status of an automatic collection task​

View metadata of statistics​

View automatic collection tasks​

Cancel a collection task​

Delete statistics​

References​

What did you think of this doc?

What is CBO

Types of statistical information

Basic statistics

Histogram

Multi-column joint statistics

Collection types and methods

Predicate Column

Collect statistics

Automatic collection

Manual collection

Manually collect basic statistics

Manually collect histograms

Custom collection

Customize an automatic collection task

Collect statistics during data loading

INSERT INTO

INSERT OVERWRITE

View custom collection tasks

Delete a custom collection task

View status of collection tasks

View statistics

View metadata of basic statistics

View metadata of histograms

Delete statistics

Delete basic statistics

Delete histograms

Cancel a collection task

Other FE configuration items

Session variable

Collect statistics of external tables

Limits

Query-triggered collection

System variable

enable_query_trigger_analyze

FE configuration

connector_table_query_trigger_analyze_small_table_rows

connector_table_query_trigger_analyze_small_table_interval

connector_table_query_trigger_analyze_large_table_interval

connector_table_query_trigger_analyze_max_pending_task_num

connector_table_query_trigger_analyze_schedule_interval

connector_table_query_trigger_analyze_max_running_task_num

Manual collection

Create a manual collection task

View the task status

View metadata of statistics

Cancel a collection task

Automatic collection

Create an automatic collection task

View the status of an automatic collection task

View metadata of statistics

View automatic collection tasks

Cancel a collection task

Delete statistics

References