文档首页/ 云数据库 GaussDB/ 特性指南（主备版_3.x）/ 分区表/ 分区表介绍/ 分区策略/ 二级分区
更新时间：2024-05-20 GMT+08:00
查看PDF
分享

二级分区

二级分区（Sub Partition，也叫组合分区）是基本数据分区类型的组合，将表通过一种数据分布方法进行分区，然后使用第二种数据分布方式将每个分区进一步细分为子分区。给定分区的所有子分区表示数据的逻辑子集。常见的二级分区组合如下所示：
  1. Range-Range
  2. Range-List
  3. Range-Hash
  4. List-Range
  5. List-List
  6. List-Hash
  7. Hash-Range
  8. Hash-List
  9. Hash-Hash

示例如下：

--Range-Range
gaussdb=# CREATE TABLE t_range_range (
    c1 INT,
    c2 INT,
    c3 INT
)
PARTITION BY RANGE (c1) 
SUBPARTITION BY RANGE (c2) 
(
    PARTITION p1 VALUES LESS THAN (10) (
        SUBPARTITION p1sp1 VALUES LESS THAN (5),
        SUBPARTITION p1sp2 VALUES LESS THAN (10)
    ),
    PARTITION p2 VALUES LESS THAN (20) (
        SUBPARTITION p2sp1 VALUES LESS THAN (15),
        SUBPARTITION p2sp2 VALUES LESS THAN (20)
    )
);
gaussdb=# DROP TABLE t_range_range;

--Range-List
gaussdb=# CREATE TABLE t_range_list (
    c1 INT,
    c2 INT,
    c3 INT
)
PARTITION BY RANGE (c1) 
SUBPARTITION BY LIST (c2) 
(
    PARTITION p1 VALUES LESS THAN (10) (
        SUBPARTITION p1sp1 VALUES (1, 2),
        SUBPARTITION p1sp2 VALUES (3, 4)
    ),
    PARTITION p2 VALUES LESS THAN (20) (
        SUBPARTITION p2sp1 VALUES (1, 2),
        SUBPARTITION p2sp2 VALUES (3, 4)
    )
);
gaussdb=# DROP TABLE t_range_list;

--Range-Hash
gaussdb=# CREATE TABLE t_range_hash (
    c1 INT,
    c2 INT,
    c3 INT
)
PARTITION BY RANGE (c1) 
SUBPARTITION BY HASH (c2) 
SUBPARTITIONS 2
(
    PARTITION p1 VALUES LESS THAN (10),
    PARTITION p2 VALUES LESS THAN (20)
);
gaussdb=# DROP TABLE t_range_hash;

--List-Range
gaussdb=# CREATE TABLE t_list_range (
    c1 INT,
    c2 INT,
    c3 INT
)
PARTITION BY LIST (c1) 
SUBPARTITION BY RANGE (c2) 
(
    PARTITION p1 VALUES (1, 2) (
        SUBPARTITION p1sp1 VALUES LESS THAN (5),
        SUBPARTITION p1sp2 VALUES LESS THAN (10)
    ),
    PARTITION p2 VALUES (3, 4) (
        SUBPARTITION p2sp1 VALUES LESS THAN (5),
        SUBPARTITION p2sp2 VALUES LESS THAN (10)
    )
);
gaussdb=# DROP TABLE t_list_range;

--List-List
gaussdb=# CREATE TABLE t_list_list (
    c1 INT,
    c2 INT,
    c3 INT
)
PARTITION BY LIST (c1) 
SUBPARTITION BY LIST (c2) 
(
    PARTITION p1 VALUES (1, 2) (
        SUBPARTITION p1sp1 VALUES (1, 2),
        SUBPARTITION p1sp2 VALUES (3, 4)
    ),
    PARTITION p2 VALUES (3, 4) (
        SUBPARTITION p2sp1 VALUES (1, 2),
        SUBPARTITION p2sp2 VALUES (3, 4)
    )
);
gaussdb=# DROP TABLE t_list_list;

--List-Hash
gaussdb=# CREATE TABLE t_list_hash (
    c1 INT,
    c2 INT,
    c3 INT
)
PARTITION BY LIST (c1) 
SUBPARTITION BY HASH (c2) 
SUBPARTITIONS 2
(
    PARTITION p1 VALUES (1, 2),
    PARTITION p2 VALUES (3, 4)
);
gaussdb=# DROP TABLE t_list_hash;

--Hash-Range
gaussdb=# CREATE TABLE t_hash_range (
    c1 INT,
    c2 INT,
    c3 INT
)
PARTITION BY HASH (c1) 
PARTITIONS 2
SUBPARTITION BY RANGE (c2) 
(
    PARTITION p1 (
        SUBPARTITION p1sp1 VALUES LESS THAN (5),
        SUBPARTITION p1sp2 VALUES LESS THAN (10)
    ),
    PARTITION p2 (
        SUBPARTITION p2sp1 VALUES LESS THAN (5),
        SUBPARTITION p2sp2 VALUES LESS THAN (10)
    )
);
gaussdb=# DROP TABLE t_hash_range;

--Hash-List
gaussdb=# CREATE TABLE t_hash_list (
    c1 INT,
    c2 INT,
    c3 INT
)
PARTITION BY HASH (c1) 
PARTITIONS 2
SUBPARTITION BY LIST (c2) 
(
    PARTITION p1 (
        SUBPARTITION p1sp1 VALUES (1, 2),
        SUBPARTITION p1sp2 VALUES (3, 4)
    ),
    PARTITION p2 (
        SUBPARTITION p2sp1 VALUES (1, 2),
        SUBPARTITION p2sp2 VALUES (3, 4)
    )
);
gaussdb=# DROP TABLE t_hash_list;

--Hash-Hash
gaussdb=# CREATE TABLE t_hash_hash (
    c1 INT,
    c2 INT,
    c3 INT
)
PARTITION BY HASH (c1) 
PARTITIONS 2
SUBPARTITION BY HASH (c2) 
SUBPARTITIONS 2
(
    PARTITION p1,
    PARTITION p2
);
gaussdb=# DROP TABLE t_hash_hash;

Interval分区看成是范围分区的一种特殊形式，目前不支持二级分区场景中定义Interval分区。

二级分区表的一级分区和二级分区分区键均只支持1列。

父主题： 分区策略