Documentation Home
MySQL 9.1 Reference Manual
Related Documentation Download this Manual
PDF (US Ltr) - 40.3Mb
PDF (A4) - 40.4Mb
Man Pages (TGZ) - 259.3Kb
Man Pages (Zip) - 366.3Kb
Info (Gzip) - 4.0Mb
Info (Zip) - 4.0Mb


15.2.12 REPLACE Statement

REPLACE [LOW_PRIORITY | DELAYED]
    [INTO] tbl_name
    [PARTITION (partition_name [, partition_name] ...)]
    [(col_name [, col_name] ...)]
    { {VALUES | VALUE} (value_list) [, (value_list)] ...
      |
      VALUES row_constructor_list
    }

REPLACE [LOW_PRIORITY | DELAYED]
    [INTO] tbl_name
    [PARTITION (partition_name [, partition_name] ...)]
    SET assignment_list

REPLACE [LOW_PRIORITY | DELAYED]
    [INTO] tbl_name
    [PARTITION (partition_name [, partition_name] ...)]
    [(col_name [, col_name] ...)]
    {SELECT ... | TABLE table_name}

value:
    {expr | DEFAULT}

value_list:
    value [, value] ...

row_constructor_list:
    ROW(value_list)[, ROW(value_list)][, ...]

assignment:
    col_name = value

assignment_list:
    assignment [, assignment] ...

REPLACE works exactly like INSERT, except that if an old row in the table has the same value as a new row for a PRIMARY KEY or a UNIQUE index, the old row is deleted before the new row is inserted. See Section 15.2.7, “INSERT Statement”.

REPLACE is a MySQL extension to the SQL standard. It either inserts, or deletes and inserts. For another MySQL extension to standard SQL—that either inserts or updates—see Section 15.2.7.2, “INSERT ... ON DUPLICATE KEY UPDATE Statement”.

DELAYED inserts and replaces were deprecated in MySQL 5.6. In MySQL 9.1, DELAYED is not supported. The server recognizes but ignores the DELAYED keyword, handles the replace as a nondelayed replace, and generates an ER_WARN_LEGACY_SYNTAX_CONVERTED warning: REPLACE DELAYED is no longer supported. The statement was converted to REPLACE. The DELAYED keyword is scheduled for removal in a future release. release.

Note

REPLACE makes sense only if a table has a PRIMARY KEY or UNIQUE index. Otherwise, it becomes equivalent to INSERT, because there is no index to be used to determine whether a new row duplicates another.

Values for all columns are taken from the values specified in the REPLACE statement. Any missing columns are set to their default values, just as happens for INSERT. You cannot refer to values from the current row and use them in the new row. If you use an assignment such as SET col_name = col_name + 1, the reference to the column name on the right hand side is treated as DEFAULT(col_name), so the assignment is equivalent to SET col_name = DEFAULT(col_name) + 1.

You can specify the column values that REPLACE attempts to insert using VALUES ROW().

To use REPLACE, you must have both the INSERT and DELETE privileges for the table.

If a generated column is replaced explicitly, the only permitted value is DEFAULT. For information about generated columns, see Section 15.1.20.8, “CREATE TABLE and Generated Columns”.

REPLACE supports explicit partition selection using the PARTITION clause with a list of comma-separated names of partitions, subpartitions, or both. As with INSERT, if it is not possible to insert the new row into any of these partitions or subpartitions, the REPLACE statement fails with the error Found a row not matching the given partition set. For more information and examples, see Section 26.5, “Partition Selection”.

The REPLACE statement returns a count to indicate the number of rows affected. This is the sum of the rows deleted and inserted. If the count is 1 for a single-row REPLACE, a row was inserted and no rows were deleted. If the count is greater than 1, one or more old rows were deleted before the new row was inserted. It is possible for a single row to replace more than one old row if the table contains multiple unique indexes and the new row duplicates values for different old rows in different unique indexes.

The affected-rows count makes it easy to determine whether REPLACE only added a row or whether it also replaced any rows: Check whether the count is 1 (added) or greater (replaced).

If you are using the C API, the affected-rows count can be obtained using the mysql_affected_rows() function.

You cannot replace into a table and select from the same table in a subquery.

MySQL uses the following algorithm for REPLACE (and LOAD DATA ... REPLACE):

  1. Try to insert the new row into the table

  2. While the insertion fails because a duplicate-key error occurs for a primary key or unique index:

    1. Delete from the table the conflicting row that has the duplicate key value

    2. Try again to insert the new row into the table

It is possible that in the case of a duplicate-key error, a storage engine may perform the REPLACE as an update rather than a delete plus insert, but the semantics are the same. There are no user-visible effects other than a possible difference in how the storage engine increments Handler_xxx status variables.

Because the results of REPLACE ... SELECT statements depend on the ordering of rows from the SELECT and this order cannot always be guaranteed, it is possible when logging these statements for the source and the replica to diverge. For this reason, REPLACE ... SELECT statements are flagged as unsafe for statement-based replication. such statements produce a warning in the error log when using statement-based mode and are written to the binary log using the row-based format when using MIXED mode. See also Section 19.2.1.1, “Advantages and Disadvantages of Statement-Based and Row-Based Replication”.

MySQL 9.1 supports TABLE as well as SELECT with REPLACE, just as it does with INSERT. See Section 15.2.7.1, “INSERT ... SELECT Statement”, for more information and examples.

When modifying an existing table that is not partitioned to accommodate partitioning, or, when modifying the partitioning of an already partitioned table, you may consider altering the table's primary key (see Section 26.6.1, “Partitioning Keys, Primary Keys, and Unique Keys”). You should be aware that, if you do this, the results of REPLACE statements may be affected, just as they would be if you modified the primary key of a nonpartitioned table. Consider the table created by the following CREATE TABLE statement:

CREATE TABLE test (
  id INT UNSIGNED NOT NULL AUTO_INCREMENT,
  data VARCHAR(64) DEFAULT NULL,
  ts TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
  PRIMARY KEY (id)
);

When we create this table and run the statements shown in the mysql client, the result is as follows:

mysql> REPLACE INTO test VALUES (1, 'Old', '2014-08-20 18:47:00');
Query OK, 1 row affected (0.04 sec)

mysql> REPLACE INTO test VALUES (1, 'New', '2014-08-20 18:47:42');
Query OK, 2 rows affected (0.04 sec)

mysql> SELECT * FROM test;
+----+------+---------------------+
| id | data | ts                  |
+----+------+---------------------+
|  1 | New  | 2014-08-20 18:47:42 |
+----+------+---------------------+
1 row in set (0.00 sec)

Now we create a second table almost identical to the first, except that the primary key now covers 2 columns, as shown here (emphasized text):

CREATE TABLE test2 (
  id INT UNSIGNED NOT NULL AUTO_INCREMENT,
  data VARCHAR(64) DEFAULT NULL,
  ts TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
  PRIMARY KEY (id, ts)
);

When we run on test2 the same two REPLACE statements as we did on the original test table, we obtain a different result:

mysql> REPLACE INTO test2 VALUES (1, 'Old', '2014-08-20 18:47:00');
Query OK, 1 row affected (0.05 sec)

mysql> REPLACE INTO test2 VALUES (1, 'New', '2014-08-20 18:47:42');
Query OK, 1 row affected (0.06 sec)

mysql> SELECT * FROM test2;
+----+------+---------------------+
| id | data | ts                  |
+----+------+---------------------+
|  1 | Old  | 2014-08-20 18:47:00 |
|  1 | New  | 2014-08-20 18:47:42 |
+----+------+---------------------+
2 rows in set (0.00 sec)

This is due to the fact that, when run on test2, both the id and ts column values must match those of an existing row for the row to be replaced; otherwise, a row is inserted.