Documentation Home
MySQL 5.7 Reference Manual
Related Documentation Download this Manual
PDF (US Ltr) - 35.0Mb
PDF (A4) - 35.1Mb
Man Pages (TGZ) - 255.5Kb
Man Pages (Zip) - 360.4Kb
Info (Gzip) - 3.4Mb
Info (Zip) - 3.4Mb
Excerpts from this Manual

MySQL 5.7 Reference Manual  /  ...  /  KEY Partitioning

22.2.5 KEY Partitioning

Partitioning by key is similar to partitioning by hash, except that where hash partitioning employs a user-defined expression, the hashing function for key partitioning is supplied by the MySQL server. NDB Cluster uses MD5() for this purpose; for tables using other storage engines, the server employs its own internal hashing function which is based on the same algorithm as PASSWORD().

The syntax rules for CREATE TABLE ... PARTITION BY KEY are similar to those for creating a table that is partitioned by hash. The major differences are listed here:

  • KEY is used rather than HASH.

  • KEY takes only a list of zero or more column names. Any columns used as the partitioning key must comprise part or all of the table's primary key, if the table has one. Where no column name is specified as the partitioning key, the table's primary key is used, if there is one. For example, the following CREATE TABLE statement is valid in MySQL 5.7:

        name VARCHAR(20)

    If there is no primary key but there is a unique key, then the unique key is used for the partitioning key:

        id INT NOT NULL,
        name VARCHAR(20),
        UNIQUE KEY (id)

    However, if the unique key column were not defined as NOT NULL, then the previous statement would fail.

    In both of these cases, the partitioning key is the id column, even though it is not shown in the output of SHOW CREATE TABLE or in the PARTITION_EXPRESSION column of the Information Schema PARTITIONS table.

    Unlike the case with other partitioning types, columns used for partitioning by KEY are not restricted to integer or NULL values. For example, the following CREATE TABLE statement is valid:

    CREATE TABLE tm1 (
        s1 CHAR(32) PRIMARY KEY

    The preceding statement would not be valid, were a different partitioning type to be specified. (In this case, simply using PARTITION BY KEY() would also be valid and have the same effect as PARTITION BY KEY(s1), since s1 is the table's primary key.)

    For additional information about this issue, see Section 22.6, “Restrictions and Limitations on Partitioning”.

    Columns with index prefixes are not supported in partitioning keys. This means that CHAR, VARCHAR, BINARY, and VARBINARY columns can be used in a partitioning key, as long as they do not employ prefixes; because a prefix must be specified for BLOB and TEXT columns in index definitions, it is not possible to use columns of these two types in partitioning keys. In MySQL 5.7, columns using prefixes are permitted when creating, altering, or upgrading partitioned tables, even though they are not included in the table's partitioning key. This is a known issue in MySQL 5.7 which is addressed in MySQL 8.0, where this permissive behavior is deprecated, and the server displays appropriate warnings or errors when attempting to use such columns in these cases. See Column index prefixes not supported for key partitioning, for more information and examples.


    Tables using the NDB storage engine are implicitly partitioned by KEY, using the table's primary key as the partitioning key (as with other MySQL storage engines). In the event that the NDB Cluster table has no explicit primary key, the hidden primary key generated by the NDB storage engine for each NDB Cluster table is used as the partitioning key.

    If you define an explicit partitioning scheme for an NDB table, the table must have an explicit primary key, and any columns used in the partitioning expression must be part of this key. However, if the table uses an empty partitioning expression—that is, PARTITION BY KEY() with no column references—then no explicit primary key is required.

    You can observe this partitioning using the ndb_desc utility (with the -p option).


    For a key-partitioned table, you cannot execute an ALTER TABLE DROP PRIMARY KEY, as doing so generates the error ERROR 1466 (HY000): Field in list of fields for partition function not found in table. This is not an issue for NDB Cluster tables which are partitioned by KEY; in such cases, the table is reorganized using the hidden primary key as the table's new partitioning key. See Chapter 21, MySQL NDB Cluster 7.5 and NDB Cluster 7.6.

It is also possible to partition a table by linear key. Here is a simple example:

    col1 INT NOT NULL,
    col2 CHAR(5),
    col3 DATE

Using LINEAR has the same effect on KEY partitioning as it does on HASH partitioning, with the partition number being derived using a powers-of-two algorithm rather than modulo arithmetic. See Section, “LINEAR HASH Partitioning”, for a description of this algorithm and its implications.