Was this page helpful?
Your feedback about this content is important. Let us know what you think.
Additional feedback?
1500 characters remaining
Collapse the table of content
Expand the table of content



Applies To: SQL Server 2014, SQL Server 2016 Preview

Creates query optimization statistics on one or more columns of a table, an indexed view, or an external table. For most queries, the query optimizer already generates the necessary statistics for a high-quality query plan; in a few cases, you need to create additional statistics with CREATE STATISTICS or modify the query design to improve query performance.

To learn more, see Statistics.

Applies to: SQL Server (SQL Server 2008 through current version), Azure SQL Database, Azure SQL Data Warehouse Public Preview.

Topic link icon Transact-SQL Syntax Conventions

--Create statistics on an external table
CREATE STATISTICS statistics_name 
ON { table_or_indexed_view_name } ( column [ ,...n ] ) 

--Create statistics on a regular table or indexed view
CREATE STATISTICS statistics_name 
ON { table_or_indexed_view_name } ( column [ ,...n ] ) 
    [ WHERE <filter_predicate> ]
    [ WITH 
        [ [ FULLSCAN 
          | SAMPLE number { PERCENT | ROWS } 
          | STATS_STREAM = stats_stream ] ] 
        [ [ , ] NORECOMPUTE ] 
        [ [ , ] INCREMENTAL = { ON | OFF } ]
    ] ;

<filter_predicate> ::= 
    <conjunct> [AND <conjunct>]

<conjunct> ::=
    <disjunct> | <comparison>

<disjunct> ::=
        column_name IN (constant ,…)

<comparison> ::=
        column_name <comparison_op> constant

<comparison_op> ::=
    IS | IS NOT | = | <> | != | > | >= | !> | < | <= | !<


Is the name of the statistics to create.


Is the name of the table, indexed view, or external table on which to create the statistics. To create statistics on another database, specify a qualified table name.

column [ ,…n]

One or more columns to be included in the statistics. The columns should be in priority order from left to right. Only the first column is used for creating the histogram. All columns are used for cross-column correlation statistics called densities.

You can specify any column that can be specified as an index key column with the following exceptions:

  • Xml, full-text, and FILESTREAM columns cannot be specified.

  • Computed columns can be specified only if the ARITHABORT and QUOTED_IDENTIFIER database settings are ON.

  • CLR user-defined type columns can be specified if the type supports binary ordering. Computed columns defined as method invocations of a user-defined type column can be specified if the methods are marked deterministic.

WHERE <filter_predicate>

Specifies an expression for selecting a subset of rows to include when creating the statistics object. Statistics that are created with a filter predicate are called filtered statistics. The filter predicate uses simple comparison logic and cannot reference a computed column, a UDT column, a spatial data type column, or a hierarchyID data type column. Comparisons using NULL literals are not allowed with the comparison operators. Use the IS NULL and IS NOT NULL operators instead.

Here are some examples of filter predicates for the Production.BillOfMaterials table:

WHERE StartDate > '20000101' AND EndDate <= '20000630'

WHERE ComponentID IN (533, 324, 753)

WHERE StartDate IN ('20000404', '20000905') AND EndDate IS NOT NULL

For more information about filter predicates, see Create Filtered Indexes.


Compute statistics by scanning all rows. FULLSCAN and SAMPLE 100 PERCENT have the same results. FULLSCAN cannot be used with the SAMPLE option.

When omitted, SQL Server uses sampling to create the statistics, and determines the sample size that is required to create a high quality query plan


Specifies the approximate percentage or number of rows in the table or indexed view for the query optimizer to use when it creates statistics. For PERCENT, number can be from 0 through 100 and for ROWS, number can be from 0 to the total number of rows. The actual percentage or number of rows the query optimizer samples might not match the percentage or number specified. For example, the query optimizer scans all rows on a data page.

SAMPLE is useful for special cases in which the query plan, based on default sampling, is not optimal. In most situations, it is not necessary to specify SAMPLE because the query optimizer already uses sampling and determines the statistically significant sample size by default, as required to create high-quality query plans.

SAMPLE cannot be used with the FULLSCAN option. When neither SAMPLE nor FULLSCAN is specified, the query optimizer uses sampled data and computes the sample size by default.

We recommend against specifying 0 PERCENT or 0 ROWS. When 0 PERCENT or ROWS is specified, the statistics object is created but does not contain statistics data.

STATS_STREAM =stats_stream

Identified for informational purposes only. Not supported. Future compatibility is not guaranteed.


Disable the automatic statistics update option, AUTO_STATISTICS_UPDATE, for statistics_name. If this option is specified, the query optimizer will complete any in-progress statistics updates for statistics_name and disable future updates.

To re-enable statistics updates, remove the statistics with DROP STATISTICS and then run CREATE STATISTICS without the NORECOMPUTE option.


Using this option can produce suboptimal query plans. We recommend using this option sparingly, and then only by a qualified system administrator.

For more information about the AUTO_STATISTICS_UPDATE option, see ALTER DATABASE SET Options (Transact-SQL). For more information about disabling and re-enabling statistics updates, see Statistics.


When ON, the statistics created are per partition statistics. When OFF, stats are combined for all partitions. The default is OFF.

If per partition statistics are not supported an error is generated. Incremental stats are not supported for following statistics types:

  • Statistics created with indexes that are not partition-aligned with the base table.

  • Statistics created on AlwaysOn readable secondary databases.

  • Statistics created on read-only databases.

  • Statistics created on filtered indexes.

  • Statistics created on views.

  • Statistics created on internal tables.

  • Statistics created with spatial indexes or XML indexes.

Applies to: SQL Server 2014 through SQL Server 2016.

Requires one of these permissions:


  • User is the table owner

  • Membership in the db_ddladmin fixed database role

SQL Server can use tempdb to sort the sampled rows before building statistics.

When creating external table statistics, SQL Server imports the external table into a temporary SQL Server table, and then creates the statistics. For samples statistics, only the sampled rows are imported. If you have a large external table, it will be much faster to use the default sampling instead of the full scan option.

Filtered statistics can improve query performance for queries that select from well-defined subsets of data. Filtered statistics use a filter predicate in the WHERE clause to select the subset of data that is included in the statistics.

For more information about when to use CREATE STATISTICS, see Statistics.

The sys.sql_expression_dependencies catalog view tracks each column in the filtered statistics predicate as a referencing dependency. Consider the operations that you perform on table columns before creating filtered statistics because you cannot drop, rename, or alter the definition of a table column that is defined in a filtered statistics predicate.

Updating statistics is not supported on external tables. To update statistics on an external table, drop and re-create the statistics.

The following example creates the ContactMail1 statistics, using a random sample of 5 percent of the BusinessEntityID and EmailPromotion columns of the Contact table of the AdventureWorks2012database.

    ON Person.Person (BusinessEntityID, EmailPromotion)

The following example creates the ContactMail2 statistics for all rows in the BusinessEntityID and EmailPromotion columns of the Contact table and disables automatic recomputing of statistics.

    ON AdventureWorks2012.Person.Person (BusinessEntityID, EmailPromotion)

The following example creates the filtered statistics ContactPromotion1. The Database Engine samples 50 percent of the data and then selects the rows with EmailPromotion equal to 2.

USE AdventureWorks2012;
IF EXISTS (SELECT name FROM sys.stats
    WHERE name = N'ContactPromotion1'
    AND object_id = OBJECT_ID(N'Person.Person'))
DROP STATISTICS Person.Person.ContactPromotion1;
    ON Person.Person (BusinessEntityID, LastName, EmailPromotion)
WHERE EmailPromotion = 2

The only decision you need to make when you create statistics on an external table, besides providing the list of columns, is whether to create the statistics by sampling the rows or by scanning all of the rows.

Since SQL Server imports data from the external table into a temporary table to create statistics, the full scan option will take much longer. For a large table, the default sampling method is usually sufficient.

--Create statistics on an external table and use default sampling.
CREATE STATISTICS CustomerStats1 ON DimCustomer (CustomerKey, EmailAddress);

--Create statistics on an external table and scan all the rows
CREATE STATISTICS CustomerStats1 ON DimCustomer (CustomerKey, EmailAddress) WITH FULLSCAN;

Community Additions

© 2015 Microsoft