Upravit

Sdílet prostřednictvím


MERGE (Transact-SQL)

Applies to: SQL Server Azure SQL Database Azure SQL Managed Instance Azure Synapse Analytics SQL database in Microsoft Fabric

The MERGE statement runs insert, update, or delete operations on a target table from the results of a join with a source table. For example, synchronize two tables by inserting, updating, or deleting rows in one table based on differences found in the other table.

Note

For MERGE information specific to Azure Synapse Analytics, change the version selection to Azure Synapse Analytics.

Note

MERGE is now Generally Available in Synapse Dedicated SQL Pool with 10.0.17829.0 and later versions. Connect to your dedicated SQL pool (formerly SQL DW) and run SELECT @@VERSION. A pause and resume might be required to ensure your instance gets the latest version.

Tip

The conditional behavior described for the MERGE statement works best when the two tables have a complex mixture of matching characteristics. For example, inserting a row if it doesn't exist, or updating a row if it matches. When simply updating one table based on the rows of another table, improve the performance and scalability with INSERT, UPDATE, and DELETE statements. For example:

INSERT tbl_A (col, col2)
SELECT col, col2
FROM tbl_B
WHERE NOT EXISTS (SELECT col FROM tbl_A A2 WHERE A2.col = tbl_B.col);

Transact-SQL syntax conventions

Syntax

Syntax for SQL Server and Azure SQL Database:

[ WITH <common_table_expression> [,...n] ]
MERGE
    [ TOP ( expression ) [ PERCENT ] ]
    [ INTO ] <target_table> [ WITH ( <merge_hint> ) ] [ [ AS ] table_alias ]
    USING <table_source> [ [ AS ] table_alias ]
    ON <merge_search_condition>
    [ WHEN MATCHED [ AND <clause_search_condition> ]
        THEN <merge_matched> ] [ ...n ]
    [ WHEN NOT MATCHED [ BY TARGET ] [ AND <clause_search_condition> ]
        THEN <merge_not_matched> ]
    [ WHEN NOT MATCHED BY SOURCE [ AND <clause_search_condition> ]
        THEN <merge_matched> ] [ ...n ]
    [ <output_clause> ]
    [ OPTION ( <query_hint> [ ,...n ] ) ]
;

<target_table> ::=
{
    [ database_name . schema_name . | schema_name . ] [ [ AS ] target_table ]
    | @variable [ [ AS ] target_table ]
    | common_table_expression_name [ [ AS ] target_table ]
}

<merge_hint>::=
{
    { [ <table_hint_limited> [ ,...n ] ]
    [ [ , ] { INDEX ( index_val [ ,...n ] ) | INDEX = index_val }]
    }
}

<merge_search_condition> ::=
    <search_condition>

<merge_matched>::=
    { UPDATE SET <set_clause> | DELETE }

<merge_not_matched>::=
{
    INSERT [ ( column_list ) ]
        { VALUES ( values_list )
        | DEFAULT VALUES }
}

<clause_search_condition> ::=
    <search_condition>

Syntax for Azure Synapse Analytics:

[ WITH <common_table_expression> [,...n] ]
MERGE
    [ INTO ] <target_table> [ [ AS ] table_alias ]
    USING <table_source> [ [ AS ] table_alias ]
    ON <merge_search_condition>
    [ WHEN MATCHED [ AND <clause_search_condition> ]
        THEN <merge_matched> ] [ ...n ]
    [ WHEN NOT MATCHED [ BY TARGET ] [ AND <clause_search_condition> ]
        THEN <merge_not_matched> ]
    [ WHEN NOT MATCHED BY SOURCE [ AND <clause_search_condition> ]
        THEN <merge_matched> ] [ ...n ]
    [ OPTION ( <query_hint> [ ,...n ] ) ]
;  -- The semi-colon is required, or the query will return a syntax error.

<target_table> ::=
{
    [ database_name . schema_name . | schema_name . ]
  target_table
}

<merge_search_condition> ::=
    <search_condition>

<merge_matched>::=
    { UPDATE SET <set_clause> | DELETE }

<merge_not_matched>::=
{
    INSERT [ ( column_list ) ]
        VALUES ( values_list )
}

<clause_search_condition> ::=
    <search_condition>

Arguments

WITH <common_table_expression>

Specifies the temporary named result set or view, also known as common table expression, that's defined within the scope of the MERGE statement. The result set derives from a simple query and is referenced by the MERGE statement. For more information, see WITH common_table_expression (Transact-SQL).

TOP ( expression ) [ PERCENT ]

Specifies the number or percentage of affected rows. expression can be either a number or a percentage of the rows. The rows referenced in the TOP expression aren't arranged in any order. For more information, see TOP (Transact-SQL).

The TOP clause applies after the entire source table and the entire target table join and the joined rows that don't qualify for an insert, update, or delete action are removed. The TOP clause further reduces the number of joined rows to the specified value. These actions (insert, update, or delete) apply to the remaining joined rows in an unordered way. That is, there's no order in which the rows are distributed among the actions defined in the WHEN clauses. For example, specifying TOP (10) affects 10 rows. Of these rows, 7 might be updated and 3 inserted, or 1 might be deleted, 5 updated, and 4 inserted, and so on.

Without filters on the source table, the MERGE statement might perform a table scan or clustered index scan on the source table, as well as a table scan or clustered index scan of target table. Therefore, I/O performance is sometimes affected even when using the TOP clause to modify a large table by creating multiple batches. In this scenario, it's important to ensure that all successive batches target new rows.

database_name

The name of the database in which target_table is located.

schema_name

The name of the schema to which target_table belongs.

target_table

The table or view against which the data rows from <table_source> are matched based on <clause_search_condition>. target_table is the target of any insert, update, or delete operations specified by the WHEN clauses of the MERGE statement.

If target_table is a view, any actions against it must satisfy the conditions for updating views. For more information, see Modify Data Through a View.

target_table can't be a remote table. target_table can't have any rules defined on it. target_table can't be a memory-optimized table.

Hints can be specified as a <merge_hint>.

<merge_hint> isn't supported for Azure Synapse Analytics.

[ AS ] table_alias

An alternative name to reference a table for the target_table.

USING <table_source>

Specifies the data source that's matched with the data rows in target_table based on <merge_search_condition>. The result of this match dictates the actions to take by the WHEN clauses of the MERGE statement. <table_source> can be a remote table or a derived table that accesses remote tables.

<table_source> can be a derived table that uses the Transact-SQL table value constructor to construct a table by specifying multiple rows.

<table_source> can be a derived table that uses SELECT ... UNION ALL to construct a table by specifying multiple rows.

[ AS ] table_alias

An alternative name to reference a table for the table_source.

For more information about the syntax and arguments of this clause, see FROM (Transact-SQL).

ON <merge_search_condition>

Specifies the conditions on which <table_source> joins with target_table to determine where they match.

Caution

It's important to specify only the columns from the target table to use for matching purposes. That is, specify columns from the target table that are compared to the corresponding column of the source table. Don't attempt to improve query performance by filtering out rows in the target table in the ON clause; for example, such as specifying AND NOT target_table.column_x = value. Doing so can return unexpected and incorrect results.

WHEN MATCHED THEN <merge_matched>

Specifies that all rows of *target_table, which match the rows returned by <table_source> ON <merge_search_condition>, and satisfy any additional search condition, are either updated or deleted according to the <merge_matched> clause.

The MERGE statement can have, at most, two WHEN MATCHED clauses. If two clauses are specified, the first clause must be accompanied by an AND <search_condition> clause. For any given row, the second WHEN MATCHED clause is only applied if the first isn't. If there are two WHEN MATCHED clauses, one must specify an UPDATE action and one must specify a DELETE action. When UPDATE is specified in the <merge_matched> clause, and more than one row of <table_source> matches a row in target_table based on <merge_search_condition>, SQL Server returns an error. The MERGE statement can't update the same row more than once, or update and delete the same row.

WHEN NOT MATCHED [ BY TARGET ] THEN <merge_not_matched>

Specifies that a row is inserted into target_table for every row returned by <table_source> ON <merge_search_condition> that doesn't match a row in target_table, but satisfies an additional search condition, if present. The values to insert are specified by the <merge_not_matched> clause. The MERGE statement can have only one WHEN NOT MATCHED [ BY TARGET ] clause.

WHEN NOT MATCHED BY SOURCE THEN <merge_matched>

Specifies that all rows of *target_table, which don't match the rows returned by <table_source> ON <merge_search_condition>, and that satisfy any additional search condition, are updated or deleted according to the <merge_matched> clause.

The MERGE statement can have at most two WHEN NOT MATCHED BY SOURCE clauses. If two clauses are specified, then the first clause must be accompanied by an AND <clause_search_condition> clause. For any given row, the second WHEN NOT MATCHED BY SOURCE clause is only applied if the first isn't. If there are two WHEN NOT MATCHED BY SOURCE clauses, then one must specify an UPDATE action and one must specify a DELETE action. Only columns from the target table can be referenced in <clause_search_condition>.

When no rows are returned by <table_source>, columns in the source table can't be accessed. If the update or delete action specified in the <merge_matched> clause references columns in the source table, error 207 (Invalid column name) is returned. For example, the clause WHEN NOT MATCHED BY SOURCE THEN UPDATE SET TargetTable.Col1 = SourceTable.Col1 can cause the statement to fail because Col1 in the source table is inaccessible.

AND <clause_search_condition>

Specifies any valid search condition. For more information, see Search Condition (Transact-SQL).

<table_hint_limited>

Specifies one or more table hints to apply on the target table for each of the insert, update, or delete actions done by the MERGE statement. The WITH keyword and the parentheses are required.

NOLOCK and READUNCOMMITTED aren't allowed. For more information about table hints, see Table Hints (Transact-SQL).

Specifying the TABLOCK hint on a table that's the target of an INSERT statement has the same effect as specifying the TABLOCKX hint. An exclusive lock is taken on the table. When FORCESEEK is specified, it applies to the implicit instance of the target table joined with the source table.

Caution

Specifying READPAST with WHEN NOT MATCHED [ BY TARGET ] THEN INSERT can result in INSERT operations that violate UNIQUE constraints.

INDEX ( index_val [ ,...n ] )

Specifies the name or ID of one or more indexes on the target table for doing an implicit join with the source table. For more information, see Table Hints (Transact-SQL).

<output_clause>

Returns a row for every row in target_table that's updated, inserted, or deleted, in no particular order. $action can be specified in the output clause. $action is a column of type nvarchar(10) that returns one of three values for each row: INSERT, UPDATE, or DELETE, according to the action done on that row. The OUTPUT clause is the recommended way to query or count rows affected by a MERGE. For more information about the arguments and behavior of this clause, see OUTPUT Clause (Transact-SQL).

OPTION ( <query_hint> [ ,...n ] )

Specifies that optimizer hints are used to customize the way the Database Engine processes the statement. For more information, see Hints (Transact-SQL) - Query.

<merge_matched>

Specifies the update or delete action that's applied to all rows of target_table that don't match the rows returned by <table_source> ON <merge_search_condition>, and which satisfy any additional search condition.

UPDATE SET <set_clause>

Specifies the list of column or variable names to update in the target table and the values with which to update them.

For more information about the arguments of this clause, see UPDATE (Transact-SQL). Setting a variable to the same value as a column isn't supported.

DELETE

Specifies that the rows matching rows in target_table are deleted.

<merge_not_matched>

Specifies the values to insert into the target table.

( column_list )

A list of one or more columns of the target table in which to insert data. Columns must be specified as a single-part name, or else the MERGE statement fails. column_list must be enclosed in parentheses and delimited by commas.

VALUES ( values_list )

A comma-separated list of constants, variables, or expressions that return values to insert into the target table. Expressions can't contain an EXECUTE statement.

DEFAULT VALUES

Forces the inserted row to contain the default values defined for each column.

For more information about this clause, see INSERT (Transact-SQL).

<search_condition>

Specifies the search conditions to specify <merge_search_condition> or <clause_search_condition>. For more information about the arguments for this clause, see Search Condition (Transact-SQL).

<graph search pattern>

Specifies the graph match pattern. For more information about the arguments for this clause, see MATCH (Transact-SQL).

Remarks

At least one of the three MATCHED clauses must be specified, but they can be specified in any order. A variable can't be updated more than once in the same MATCHED clause.

Any insert, update, or delete action specified on the target table by the MERGE statement are limited by any constraints defined on it, including any cascading referential integrity constraints. If IGNORE_DUP_KEY is ON for any unique indexes on the target table, MERGE ignores this setting.

The MERGE statement requires a semicolon (;) as a statement terminator. Error 10713 is raised when a MERGE statement is run without the terminator.

When used after MERGE, @@ROWCOUNT (Transact-SQL) returns the total number of rows inserted, updated, and deleted to the client.

MERGE is a fully reserved keyword when the database compatibility level is set to 100 or higher. The MERGE statement is available under both 90 and 100 database compatibility levels; however, the keyword isn't fully reserved when the database compatibility level is set to 90.

Caution

Don't use the MERGE statement when using queued updating replication. The MERGE and queued updating trigger aren't compatible. Replace the MERGE statement with an insert or an update statement.

Azure Synapse Analytics considerations

In Azure Synapse Analytics, the MERGE command has following differences compared to SQL Server and Azure SQL database.

  • Using MERGE to update a distribution key column isn't supported in builds older than 10.0.17829.0. If unable to pause or force-upgrade, use the ANSI UPDATE FROM ... JOIN statement as a workaround until on version 10.0.17829.0.
  • A MERGE update is implemented as a delete and insert pair. The affected row count for a MERGE update includes the deleted and inserted rows.
  • MERGE...WHEN NOT MATCHED INSERT isn't supported for tables with IDENTITY columns.
  • Table value constructor can't be used in the USING clause for the source table. Use SELECT ... UNION ALL to create a derived source table with multiple rows.
  • The support for tables with different distribution types is described in this table:
MERGE CLAUSE in Azure Synapse Analytics Supported TARGET distribution table Supported SOURCE distribution table Comment
WHEN MATCHED All distribution types All distribution types
NOT MATCHED BY TARGET HASH All distribution types Use UPDATE/DELETE FROM...JOIN to synchronize two tables.
NOT MATCHED BY SOURCE All distribution types All distribution types

Tip

If you're using the distribution hash key as the JOIN column in MERGE and performing just an equality comparison, you can omit the distribution key from the list of columns in the WHEN MATCHED THEN UPDATE SET clause, as this is a redundant update.

In Azure Synapse Analytics, the MERGE command on builds older than 10.0.17829.0 can, under certain conditions, leave the target table in an inconsistent state, with rows placed in the wrong distribution, causing later queries to return wrong results in some cases. This problem might happen in 2 cases:

Scenario Comment
Case 1
Using MERGE on a HASH distributed TARGET table that contains secondary indices or a UNIQUE constraint.
- Fixed in Synapse SQL 10.0.15563.0 and later versions.
- If SELECT @@VERSION returns a lower version than 10.0.15563.0, manually pause and resume the Synapse SQL pool to pick up this fix.
- Until the fix has been applied to your Synapse SQL pool, avoid using the MERGE command on HASH distributed TARGET tables that have secondary indices or UNIQUE constraints.
Case 2
Using MERGE to update a distribution key column of a HASH distributed table.
- Fixed in Synapse SQL 10.0.17829.0 and later versions.
- If SELECT @@VERSION returns a lower version than 10.0.17829.0, manually pause and resume the Synapse SQL pool to pick up this fix.
- Until the fix has been applied to your Synapse SQL pool, avoid using the MERGE command to update distribution key columns.

The updates in both scenarios do not repair tables already affected by previous MERGE execution. Use the following scripts to identify and repair any affected tables manually.

To check which HASH distributed tables in a database might be of concern (if used in the previously mentioned cases), run this statement:

-- Case 1
SELECT a.name,
    c.distribution_policy_desc,
    b.type
FROM sys.tables a
INNER JOIN sys.indexes b
    ON a.object_id = b.object_id
INNER JOIN sys.pdw_table_distribution_properties c
    ON a.object_id = c.object_id
WHERE b.type = 2
    AND c.distribution_policy_desc = 'HASH';

-- Subject to Case 2, if distribution key value is updated in MERGE statement
SELECT a.name,
    c.distribution_policy_desc
FROM sys.tables a
INNER JOIN sys.pdw_table_distribution_properties c
    ON a.object_id = c.object_id
WHERE c.distribution_policy_desc = 'HASH';

To check if a HASH distributed table for MERGE is affected by either Case 1 or Case 2, follow these steps to examine if the tables have rows landed in wrong distribution. If no need for repair is returned, this table is not affected.

IF object_id('[check_table_1]', 'U') IS NOT NULL
    DROP TABLE [check_table_1]
GO

IF object_id('[check_table_2]', 'U') IS NOT NULL
    DROP TABLE [check_table_2]
GO

CREATE TABLE [check_table_1]
    WITH (DISTRIBUTION = ROUND_ROBIN) AS

SELECT <DISTRIBUTION_COLUMN> AS x
FROM <MERGE_TABLE>
GROUP BY <DISTRIBUTION_COLUMN>;
GO

CREATE TABLE [check_table_2]
    WITH (DISTRIBUTION = HASH (x)) AS

SELECT x
FROM [check_table_1];
GO

IF NOT EXISTS (
        SELECT TOP 1 *
        FROM (
            SELECT <DISTRIBUTION_COLUMN> AS x
            FROM <MERGE_TABLE>

            EXCEPT

            SELECT x
            FROM [check_table_2]
            ) AS tmp
        )
    SELECT 'no need for repair' AS result
ELSE
    SELECT 'needs repair' AS result
GO

IF object_id('[check_table_1]', 'U') IS NOT NULL
    DROP TABLE [check_table_1]
GO

IF object_id('[check_table_2]', 'U') IS NOT NULL
    DROP TABLE [check_table_2]
GO

To repair affected tables, run these statements to copy all rows from the old table to a new table.

IF object_id('[repair_table_temp]', 'U') IS NOT NULL
    DROP TABLE [repair_table_temp];
GO

IF object_id('[repair_table]', 'U') IS NOT NULL
    DROP TABLE [repair_table];
GO

CREATE TABLE [repair_table_temp]
    WITH (DISTRIBUTION = ROUND_ROBIN) AS

SELECT *
FROM <MERGE_TABLE>;
GO

-- [repair_table] will hold the repaired table generated from <MERGE_TABLE>
CREATE TABLE [repair_table]
    WITH (DISTRIBUTION = HASH (<DISTRIBUTION_COLUMN>)) AS

SELECT *
FROM [repair_table_temp];
GO

IF object_id('[repair_table_temp]', 'U') IS NOT NULL
    DROP TABLE [repair_table_temp];
GO

Troubleshooting

In certain scenarios, a MERGE statement might result in the error CREATE TABLE failed because column <> in table <> exceeds the maximum of 1024 columns., even when the target or source table don't have 1,024 columns. This scenario can arise when any of the following conditions are met:

  • Multiple columns are specified in an DELETE, UPDATE SET, or INSERT operation within MERGE (not specific to any WHEN [NOT] MATCHED clause)
  • Any column in the JOIN condition has a nonclustered index (NCI)
  • Target table is HASH distributed

If this error is found, the suggested workarounds are as follows:

  • Remove the nonclustered index (NCI) from the JOIN columns or join on columns without an NCI. If you later update the underlying tables to include an NCI on the JOIN columns, your MERGE statement might be susceptible to this error at runtime. For more information, see DROP INDEX.
  • Use UPDATE, DELETE, and INSERT statements instead of MERGE.

Trigger implementation

For every insert, update, or delete action specified in the MERGE statement, SQL Server fires any corresponding AFTER triggers defined on the target table, but doesn't guarantee on which action to fire triggers first or last. Triggers defined for the same action honor the order you specify. For more information about setting trigger firing order, see Specify First and Last Triggers.

If the target table has an enabled INSTEAD OF trigger defined on it for an insert, update, or delete action done by a MERGE statement, it must have an enabled INSTEAD OF trigger for all of the actions specified in the MERGE statement.

If any INSTEAD OF UPDATE or INSTEAD OF DELETE triggers are defined on target_table, the update or delete operations aren't run. Instead, the triggers fire and the inserted and deleted tables then populate accordingly.

If any INSTEAD OF INSERT triggers are defined on target_table, the insert operation isn't performed. Instead, the table populates accordingly.

Note

Unlike separate INSERT, UPDATE, and DELETE statements, the number of rows reflected by @@ROWCOUNT inside of a trigger might be higher. The @@ROWCOUNT inside any AFTER trigger (regardless of data modification statements the trigger captures) will reflect the total number of rows affected by the MERGE. For example, if a MERGE statement inserts one row, updates one row, and deletes one row, @@ROWCOUNT will be three for any AFTER trigger, even if the trigger is only declared for INSERT statements.

Permissions

Requires SELECT permission on the source table and INSERT, UPDATE, or DELETE permissions on the target table. For more information, see the Permissions section in the SELECT, INSERT, UPDATE, and DELETE articles.

Index best practices

By using the MERGE statement, you can replace the individual DML statements with a single statement. This can improve query performance because the operations are performed within a single statement, therefore, minimizing the number of times the data in the source and target tables are processed. However, performance gains depend on having correct indexes, joins, and other considerations in place.

To improve the performance of the MERGE statement, we recommend the following index guidelines:

  • Create indexes to facilitate the join between the source and target of the MERGE:
    • Create an index on the join columns in the source table that has keys covering the join logic to the target table. If possible, it should be unique.
    • Also, create an index on the join columns in the target table. If possible, it should be a unique clustered index.
    • These two indexes ensure that the data in the tables is sorted, and uniqueness aids performance of the comparison. Query performance is improved because the query optimizer doesn't need to perform extra validation processing to locate and update duplicate rows and additional sort operations aren't necessary.
  • Avoid tables with any form of columnstore index as the target of MERGE statements. As with any UPDATEs, you might find performance better with columnstore indexes by updating a staged rowstore table, then performing a batched DELETE and INSERT, instead of an UPDATE or MERGE.

Concurrency considerations for MERGE

In terms of locking, MERGE is different from discrete, consecutive INSERT, UPDATE, and DELETE statements. MERGE still executes INSERT, UPDATE, and DELETE operations, however using different locking mechanisms. It might be more efficient to write discrete INSERT, UPDATE, and DELETE statements for some application needs. At scale, MERGE might introduce complicated concurrency issues or require advanced troubleshooting. As such, plan to thoroughly test any MERGE statement before deploying to production.

MERGE statements are a suitable replacement for discrete INSERT, UPDATE, and DELETE operations in (but not limited to) the following scenarios:

  • ETL operations involving large row counts be executed during a time when other concurrent operations aren't* expected. When heavy concurrency is expected, separate INSERT, UPDATE, and DELETE logic might perform better, with less blocking, than a MERGE statement.
  • Complex operations involving small row counts and transactions unlikely to execute for extended duration.
  • Complex operations involving user tables where indexes can be designed to ensure optimal execution plans, avoiding table scans and lookups in favor of index scans or - ideally - index seeks.

Other considerations for concurrency:

  • In some scenarios where unique keys are expected to be both inserted and updated by the MERGE, specifying the HOLDLOCK will prevent against unique key violations. HOLDLOCK is a synonym for the SERIALIZABLE transaction isolation level, which doesn't allow for other concurrent transactions to modify data that this transaction has read. SERIALIZABLE is the safest isolation level but provides for the least concurrency with other transactions that retains locks on ranges of data to prevent phantom rows from being inserted or updated while reads are in progress. For more information on HOLDLOCK, see Hints and SET TRANSACTION ISOLATION LEVEL (Transact-SQL).

JOIN best practices

To improve the performance of the MERGE statement and ensure correct results are obtained, we recommend the following join guidelines:

  • Specify only search conditions in the ON <merge_search_condition> clause that determine the criteria for matching data in the source and target tables. That is, specify only columns from the target table that are compared to the corresponding columns of the source table.
  • Don't include comparisons to other values such as a constant.

To filter out rows from the source or target tables, use one of the following methods.

  • Specify the search condition for row filtering in the appropriate WHEN clause. For example, WHEN NOT MATCHED AND S.EmployeeName LIKE 'S%' THEN INSERT....
  • Define a view on the source or target that returns the filtered rows and reference the view as the source or target table. If the view is defined on the target table, any actions against it must satisfy the conditions for updating views. For more information about updating data by using a view, see Modifying Data Through a View.
  • Use the WITH <common table expression> clause to filter out rows from the source or target tables. This method is similar to specifying additional search criteria in the ON clause and might produce incorrect results. We recommend that you avoid using this method or test thoroughly before implementing it.

The join operation in the MERGE statement is optimized in the same way as a join in a SELECT statement. That is, when SQL Server processes join, the query optimizer chooses the most efficient method (out of several possibilities) of processing the join. When the source and target are of similar size and the index guidelines described previously are applied to the source and target tables, a merge join operator is the most efficient query plan. This is because both tables are scanned once and there's no need to sort the data. When the source is smaller than the target table, a nested loops operator is preferable.

You can force the use of a specific join by specifying the OPTION (<query_hint>) clause in the MERGE statement. We recommend that you don't use the hash join as a query hint for MERGE statements because this join type doesn't use indexes.

Parameterization best practices

If a SELECT, INSERT, UPDATE, or DELETE statement is executed without parameters, the SQL Server query optimizer might choose to parameterize the statement internally. This means that any literal values that are contained in the query are substituted with parameters. For example, the statement INSERT dbo.MyTable (Col1, Col2) VALUES (1, 10), might be implemented internally as INSERT dbo.MyTable (Col1, Col2) VALUES (@p1, @p2). This process, which is called simple parameterization, increases the ability of the relational engine to match new SQL statements with existing, previously compiled execution plans. Query performance might be improved because the frequency of query compilations and recompilations is reduced. The query optimizer doesn't apply the simple parameterization process to MERGE statements. Therefore, MERGE statements that contain literal values might not perform and individual INSERT, UPDATE, or DELETE statements because a new plan is compiled each time the MERGE statement is executed.

To improve query performance, we recommend the following parameterization guidelines:

  • Parameterize all literal values in the ON <merge_search_condition> clause and in the WHEN clauses of the MERGE statement. For example, you can incorporate the MERGE statement into a stored procedure replacing the literal values with appropriate input parameters.
  • If you can't parameterize the statement, create a plan guide of type TEMPLATE and specify the PARAMETERIZATION FORCED query hint in the plan guide. For more information, see Specify Query Parameterization Behavior by Using Plan Guides.
  • If MERGE statements are executed frequently on the database, consider setting the PARAMETERIZATION option on the database to FORCED. Use caution when setting this option. The PARAMETERIZATION option is a database-level setting and affects how all queries against the database are processed. For more information, see Forced Parameterization.
  • As a newer and easier alternative to plan guides, consider a similar strategy with Query Store hints. For more information, see Query Store hints.

TOP clause best practices

In the MERGE statement, the TOP clause specifies the number or percentage of rows that are affected after the source table and the target table are joined, and after rows that don't qualify for an insert, update, or delete action are removed. The TOP clause further reduces the number of joined rows to the specified value and the insert, update, or delete actions are applied to the remaining joined rows in an unordered fashion. That is, there's no order in which the rows are distributed among the actions defined in the WHEN clauses. For example, specifying TOP (10) affects 10 rows; of these rows, 7 might be updated and 3 inserted, or 1 might be deleted, 5 updated, and 4 inserted and so on.

It's common to use the TOP clause to perform data manipulation language (DML) operations on a large table in batches. When using the TOP clause in the MERGE statement for this purpose, it's important to understand the following implications.

  • I/O performance might be affected.

    The MERGE statement performs a full table scan of both the source and target tables. Dividing the operation into batches reduces the number of write operations performed per batch; however, each batch performs a full table scan of the source and target tables. The resulting read activity might affect the performance of the query and other concurrent activity on the tables.

  • Incorrect results can occur.

    It's important to ensure that all successive batches target new rows or undesired behavior such as incorrectly inserting duplicate rows into the target table can occur. This can happen when the source table includes a row that wasn't in a target batch but was in the overall target table. To ensure correct results:

    • Use the ON clause to determine which source rows affect existing target rows and which are genuinely new.
    • Use an additional condition in the WHEN MATCHED clause to determine if the target row was already updated by a previous batch.
    • Use an additional condition in the WHEN MATCHED clause and SET logic to verify the same row can't be updated twice.

Because the TOP clause is only applied after these clauses are applied, each execution either inserts one genuinely unmatched row or updates one existing row.

Bulk load best practices

The MERGE statement can be used to efficiently bulk load data from a source data file into a target table by specifying the OPENROWSET(BULK...) clause as the table source. By doing so, the entire file is processed in a single batch.

To improve the performance of the bulk merge process, we recommend the following guidelines:

  • Create a clustered index on the join columns in the target table.

  • Disable other non-unique, nonclustered indexes on the target table during the bulk load MERGE, enable them afterwards. This is common and useful for nightly bulk data operations.

  • Use the ORDER and UNIQUE hints in the OPENROWSET(BULK...) clause, to specify how the source data file is sorted.

    By default, the bulk operation assumes the data file is unordered. Therefore, it's important that the source data is sorted according to the clustered index on the target table and that the ORDER hint is used to indicate the order so that the query optimizer can generate a more efficient query plan. Hints are validated at runtime; if the data stream doesn't conform to the specified hints, an error is raised.

These guidelines ensure that the join keys are unique and the sort order of the data in the source file matches the target table. Query performance is improved because additional sort operations aren't necessary and unnecessary data copies aren't required.

Measure and diagnose MERGE performance

The following features are available to assist you in measuring and diagnosing the performance of MERGE statements.

Examples

A. Use MERGE to do INSERT and UPDATE operations on a table in a single statement

A common scenario is updating one or more columns in a table if a matching row exists. Or, inserting the data as a new row if a matching row doesn't exist. You usually do either scenario by passing parameters to a stored procedure that contains the appropriate UPDATE and INSERT statements. With the MERGE statement, you can do both tasks in a single statement. The following example shows a stored procedure in the AdventureWorks2022 database that contains both an INSERT statement and an UPDATE statement. The procedure is then modified to run the equivalent operations by using a single MERGE statement.

CREATE PROCEDURE dbo.InsertUnitMeasure @UnitMeasureCode NCHAR(3), @Name NVARCHAR(25)
AS
BEGIN
    SET NOCOUNT ON;

    -- Update the row if it exists.
    UPDATE Production.UnitMeasure
    SET Name = @Name
    WHERE UnitMeasureCode = @UnitMeasureCode

    -- Insert the row if the UPDATE statement failed.
    IF (@@ROWCOUNT = 0)
    BEGIN
        INSERT INTO Production.UnitMeasure (
            UnitMeasureCode,
            Name
        )
        VALUES (@UnitMeasureCode, @Name)
    END
END;
GO

-- Test the procedure and return the results.
EXEC InsertUnitMeasure @UnitMeasureCode = 'ABC', @Name = 'Test Value';

SELECT UnitMeasureCode, Name
FROM Production.UnitMeasure
WHERE UnitMeasureCode = 'ABC';
GO

-- Rewrite the procedure to perform the same operations using the
-- MERGE statement.
-- Create a temporary table to hold the updated or inserted values
-- from the OUTPUT clause.
CREATE TABLE #MyTempTable (
    ExistingCode NCHAR(3),
    ExistingName NVARCHAR(50),
    ExistingDate DATETIME,
    ActionTaken NVARCHAR(10),
    NewCode NCHAR(3),
    NewName NVARCHAR(50),
    NewDate DATETIME
);
GO

ALTER PROCEDURE dbo.InsertUnitMeasure @UnitMeasureCode NCHAR(3),
    @Name NVARCHAR(25)
AS
BEGIN
    SET NOCOUNT ON;

    MERGE Production.UnitMeasure AS tgt
    USING (SELECT @UnitMeasureCode, @Name) AS src(UnitMeasureCode, Name)
        ON (tgt.UnitMeasureCode = src.UnitMeasureCode)
    WHEN MATCHED
        THEN
            UPDATE
            SET Name = src.Name
    WHEN NOT MATCHED
        THEN
            INSERT (UnitMeasureCode, Name)
            VALUES (src.UnitMeasureCode, src.Name)
    OUTPUT deleted.*,
        $action,
        inserted.*
    INTO #MyTempTable;
END;
GO

-- Test the procedure and return the results.
EXEC InsertUnitMeasure @UnitMeasureCode = 'ABC', @Name = 'New Test Value';
EXEC InsertUnitMeasure @UnitMeasureCode = 'XYZ', @Name = 'Test Value';
EXEC InsertUnitMeasure @UnitMeasureCode = 'ABC', @Name = 'Another Test Value';

SELECT * FROM #MyTempTable;

-- Cleanup
DELETE FROM Production.UnitMeasure
WHERE UnitMeasureCode IN ('ABC', 'XYZ');

DROP TABLE #MyTempTable;
GO
CREATE PROCEDURE dbo.InsertUnitMeasure @UnitMeasureCode NCHAR(3),
    @Name NVARCHAR(25)
AS
BEGIN
    SET NOCOUNT ON;

    -- Update the row if it exists.
    UPDATE Production.UnitMeasure
    SET Name = @Name
    WHERE UnitMeasureCode = @UnitMeasureCode

    -- Insert the row if the UPDATE statement failed.
    IF (@@ROWCOUNT = 0)
    BEGIN
        INSERT INTO Production.UnitMeasure (
            UnitMeasureCode,
            Name
        )
        VALUES (@UnitMeasureCode, @Name)
    END
END;
GO

-- Test the procedure and return the results.
EXEC InsertUnitMeasure @UnitMeasureCode = 'ABC', @Name = 'Test Value';

SELECT UnitMeasureCode, Name
FROM Production.UnitMeasure
WHERE UnitMeasureCode = 'ABC';
GO

-- Rewrite the procedure to perform the same operations using the
-- MERGE statement.
ALTER PROCEDURE dbo.InsertUnitMeasure @UnitMeasureCode NCHAR(3),
    @Name NVARCHAR(25)
AS
BEGIN
    SET NOCOUNT ON;

    MERGE Production.UnitMeasure AS tgt
    USING (
        SELECT @UnitMeasureCode,
            @Name
        ) AS src(UnitMeasureCode, Name)
        ON (tgt.UnitMeasureCode = src.UnitMeasureCode)
    WHEN MATCHED
        THEN
            UPDATE SET Name = src.Name
    WHEN NOT MATCHED
        THEN
            INSERT (UnitMeasureCode, Name)
            VALUES (src.UnitMeasureCode, src.Name);
END;
GO

-- Test the procedure and return the results.
EXEC InsertUnitMeasure @UnitMeasureCode = 'ABC', @Name = 'New Test Value';
EXEC InsertUnitMeasure @UnitMeasureCode = 'XYZ', @Name = 'Test Value';
EXEC InsertUnitMeasure @UnitMeasureCode = 'ABC', @Name = 'Another Test Value';

-- Cleanup
DELETE FROM Production.UnitMeasure
WHERE UnitMeasureCode IN ('ABC', 'XYZ');
GO

B. Use MERGE to do UPDATE and DELETE operations on a table in a single statement

The following example uses MERGE to update the ProductInventory table in the AdventureWorks2022 sample database, daily, based on orders that are processed in the SalesOrderDetail table. The Quantity column of the ProductInventory table is updated by subtracting the number of orders placed each day for each product in the SalesOrderDetail table. If the number of orders for a product drops the inventory level of a product to 0 or less, the row for that product is deleted from the ProductInventory table.

CREATE PROCEDURE Production.usp_UpdateInventory @OrderDate DATETIME
AS
MERGE Production.ProductInventory AS tgt
USING (
    SELECT ProductID,
        SUM(OrderQty)
    FROM Sales.SalesOrderDetail AS sod
    INNER JOIN Sales.SalesOrderHeader AS soh
        ON sod.SalesOrderID = soh.SalesOrderID
            AND soh.OrderDate = @OrderDate
    GROUP BY ProductID
    ) AS src(ProductID, OrderQty)
    ON (tgt.ProductID = src.ProductID)
WHEN MATCHED
    AND tgt.Quantity - src.OrderQty <= 0
    THEN
        DELETE
WHEN MATCHED
    THEN
        UPDATE
        SET tgt.Quantity = tgt.Quantity - src.OrderQty,
            tgt.ModifiedDate = GETDATE()
OUTPUT $action,
    Inserted.ProductID,
    Inserted.Quantity,
    Inserted.ModifiedDate,
    Deleted.ProductID,
    Deleted.Quantity,
    Deleted.ModifiedDate;
GO

EXECUTE Production.usp_UpdateInventory '20030501';
CREATE PROCEDURE Production.usp_UpdateInventory @OrderDate DATETIME
AS
MERGE Production.ProductInventory AS tgt
USING (
    SELECT ProductID,
        SUM(OrderQty)
    FROM Sales.SalesOrderDetail AS sod
    INNER JOIN Sales.SalesOrderHeader AS soh
        ON sod.SalesOrderID = soh.SalesOrderID
            AND soh.OrderDate = @OrderDate
    GROUP BY ProductID
    ) AS src(ProductID, OrderQty)
    ON (tgt.ProductID = src.ProductID)
WHEN MATCHED
    AND tgt.Quantity - src.OrderQty <= 0
    THEN
        DELETE
WHEN MATCHED
    THEN
        UPDATE
        SET tgt.Quantity = tgt.Quantity - src.OrderQty,
            tgt.ModifiedDate = GETDATE();
GO

EXECUTE Production.usp_UpdateInventory '20030501';

C. Use MERGE to do UPDATE and INSERT operations on a target table by using a derived source table

The following example uses MERGE to modify the SalesReason table in the AdventureWorks2022 database by either updating or inserting rows.

When the value of NewName in the source table matches a value in the Name column of the target table, (SalesReason), the ReasonType column is updated in the target table. When the value of NewName doesn't match, the source row is inserted into the target table. The source table is a derived table that uses the Transact-SQL table value constructor to specify multiple rows for the source table. For more information about using the table value constructor in a derived table, see Table Value Constructor (Transact-SQL).

The OUTPUT clause can be useful to query the result of MERGE statements, for more information, see OUTPUT Clause. The example also shows how to store the results of the OUTPUT clause in a table variable. And, then you summarize the results of the MERGE statement by running a simple select operation that returns the count of inserted and updated rows.

-- Create a temporary table variable to hold the output actions.
DECLARE @SummaryOfChanges TABLE (Change VARCHAR(20));

MERGE INTO Sales.SalesReason AS tgt
USING (
    VALUES ('Recommendation', 'Other'),
        ('Review', 'Marketing'),
        ('Internet', 'Promotion')
    ) AS src(NewName, NewReasonType)
    ON tgt.Name = src.NewName
WHEN MATCHED
    THEN
        UPDATE
        SET ReasonType = src.NewReasonType
WHEN NOT MATCHED BY TARGET
    THEN
        INSERT (Name, ReasonType)
        VALUES (NewName, NewReasonType)
OUTPUT $action
INTO @SummaryOfChanges;

-- Query the results of the table variable.
SELECT Change,
    COUNT(*) AS CountPerChange
FROM @SummaryOfChanges
GROUP BY Change;

When the value of NewName in the source table matches a value in the Name column of the target table, (SalesReason), the ReasonType column is updated in the target table. When the value of NewName doesn't match, the source row is inserted into the target table. The source table is a derived table that uses SELECT ... UNION ALL to specify multiple rows for the source table.

MERGE INTO Sales.SalesReason AS tgt
USING (
    SELECT 'Recommendation', 'Other'
    UNION ALL    
    SELECT 'Review', 'Marketing'
    UNION ALL
    SELECT 'Internet', 'Promotion'
    ) AS src(NewName, NewReasonType)
    ON tgt.Name = src.NewName
WHEN MATCHED
    THEN
        UPDATE SET ReasonType = src.NewReasonType
WHEN NOT MATCHED BY TARGET
    THEN
        INSERT (Name, ReasonType)
        VALUES (NewName, NewReasonType);

D. Insert the results of the MERGE statement into another table

The following example captures data returned from the OUTPUT clause of a MERGE statement and inserts that data into another table. The MERGE statement updates the Quantity column of the ProductInventory table in the AdventureWorks2022 database, based on orders that are processed in the SalesOrderDetail table. The example captures the updated rows and inserts them into another table that's used to track inventory changes.

CREATE TABLE Production.UpdatedInventory (
    ProductID INT NOT NULL,
    LocationID INT,
    NewQty INT,
    PreviousQty INT,
    CONSTRAINT PK_Inventory PRIMARY KEY CLUSTERED (
        ProductID,
        LocationID
        )
    );
GO

INSERT INTO Production.UpdatedInventory
SELECT ProductID, LocationID, NewQty, PreviousQty
FROM (
    MERGE Production.ProductInventory AS pi
    USING (
        SELECT ProductID, SUM(OrderQty)
        FROM Sales.SalesOrderDetail AS sod
        INNER JOIN Sales.SalesOrderHeader AS soh
            ON sod.SalesOrderID = soh.SalesOrderID
                AND soh.OrderDate BETWEEN '20030701'
                    AND '20030731'
        GROUP BY ProductID
        ) AS src(ProductID, OrderQty)
        ON pi.ProductID = src.ProductID
    WHEN MATCHED
        AND pi.Quantity - src.OrderQty >= 0
        THEN
            UPDATE SET pi.Quantity = pi.Quantity - src.OrderQty
    WHEN MATCHED
        AND pi.Quantity - src.OrderQty <= 0
        THEN
            DELETE
    OUTPUT $action,
        Inserted.ProductID,
        Inserted.LocationID,
        Inserted.Quantity AS NewQty,
        Deleted.Quantity AS PreviousQty
    ) AS Changes(Action, ProductID, LocationID, NewQty, PreviousQty)
WHERE Action = 'UPDATE';
GO

E. Use MERGE to do INSERT or UPDATE on a target edge table in a graph database

In this example, you create node tables Person and City and an edge table livesIn. You use the MERGE statement on the livesIn edge and insert a new row if the edge doesn't already exist between a Person and City. If the edge already exists, then you just update the StreetAddress attribute on the livesIn edge.

-- CREATE node and edge tables
CREATE TABLE Person
(
    ID INTEGER PRIMARY KEY,
    PersonName VARCHAR(100)
)
AS NODE
GO

CREATE TABLE City
(
    ID INTEGER PRIMARY KEY,
    CityName VARCHAR(100),
    StateName VARCHAR(100)
)
AS NODE
GO

CREATE TABLE livesIn
(
    StreetAddress VARCHAR(100)
)
AS EDGE
GO

-- INSERT some test data into node and edge tables
INSERT INTO Person VALUES (1, 'Ron'), (2, 'David'), (3, 'Nancy')
GO

INSERT INTO City VALUES (1, 'Redmond', 'Washington'), (2, 'Seattle', 'Washington')
GO

INSERT livesIn SELECT P.$node_id, C.$node_id, c
FROM Person P, City C, (values (1,1, '123 Avenue'), (2,2,'Main Street')) v(a,b,c)
WHERE P.id = a AND C.id = b
GO

-- Use MERGE to update/insert edge data
CREATE OR ALTER PROCEDURE mergeEdge
    @PersonId integer,
    @CityId integer,
    @StreetAddress varchar(100)
AS
BEGIN
    MERGE livesIn
        USING ((SELECT @PersonId, @CityId, @StreetAddress) AS T (PersonId, CityId, StreetAddress)
                JOIN Person ON T.PersonId = Person.ID
                JOIN City ON T.CityId = City.ID)
        ON MATCH (Person-(livesIn)->City)
    WHEN MATCHED THEN
        UPDATE SET StreetAddress = @StreetAddress
    WHEN NOT MATCHED THEN
        INSERT ($from_id, $to_id, StreetAddress)
        VALUES (Person.$node_id, City.$node_id, @StreetAddress) ;
END
GO

-- Following will insert a new edge in the livesIn edge table
EXEC mergeEdge 3, 2, '4444th Avenue'
GO

-- Following will update the StreetAddress on the edge that connects Ron to Redmond
EXEC mergeEdge 1, 1, '321 Avenue'
GO

-- Verify that all the address were added/updated correctly
SELECT PersonName, CityName, StreetAddress
FROM Person , City , livesIn
WHERE MATCH(Person-(livesIn)->city)
GO