功能

MERGE INTO
MERGE INTO
语句用于根据源表或子查询中的值更新目标表中的记录。当源表包含目标表中的新行（待插入）、修改行（待更新）和删除行（待删除）时，可以使用此功能来同步目标表的数据。

语法

MERGE INTO target_table USING source_table ON merge_condition { WHEN MATCHED [ AND matched_condition ] THEN matched_action | WHEN MATCHED [ AND matched_condition ] THEN matched_action | --当WHEN MATCHED语句和WHEN NOT MATCHED 语句同时存在时，MERGE INTO语句中只能存在一个WHEN NOT MATCHED WHEN NOT MATCHED [ AND not_matched_by_source_condition ] THEN not_matched_by_source_action } -- 参数解释 matched_action ::= UPDATE SET <column_name> = <expr> [ , <column_name2> = <expr2> ... ] | DELETE not_matched_action ::= INSERT [ ( <column_name> [ , ... ] ) ] VALUES ( <expr> [ , ... ] )

参数说明

  • target_table
    target_table
    ：指定目标表，可以使用别名。
  • source_table
    source_table
    ：指定要与目标表进行比较的表或子查询，可以使用别名。
  • merge_condition
    merge_condition
    ：定义如何将源表中的行与目标表中的行进行匹配，返回布尔类型的表达式。
  • WHEN MATCHED [ AND matched_condition ]
    WHEN MATCHED [ AND matched_condition ]
    ：当源行根据 
    merge_condition
    merge_condition
    和可选的 
    matched_condition
    matched_condition
    与目标表行匹配时，执行 
    WHEN MATCHED
    WHEN MATCHED
    子句。
    • 如果对于匹配 
      merge_condition
      merge_condition
      的源行和目标行对，没有任何 
      WHEN MATCHED
      WHEN MATCHED
      条件的计算结果为 true，则目标行保持不变。
    • 如果存在多个 
      WHEN MATCHED
      WHEN MATCHED
      子句，则会按照它们的指定顺序对其进行求值。每个 
      WHEN MATCHED
      WHEN MATCHED
      子句都必须具有一个 
      matched_condition
      matched_condition
      。
  • WHEN NOT MATCHED [ AND not_matched_condition ]
    WHEN NOT MATCHED [ AND not_matched_condition ]
    ：当源行根据 
    merge_condition
    merge_condition
    和可选的 
    not_matched_condition
    not_matched_condition
    与目标表中的任何行都不匹配时，执行 
    WHEN NOT MATCHED
    WHEN NOT MATCHED
    子句。
    • 只能存在一个
      WHEN NOT MATCHED
      WHEN NOT MATCHED
      语句
  • matched_action
    matched_action
    ： 
    • DELETE
      DELETE
      ：删除匹配的目标表行。
    • UPDATE
      UPDATE
      ：更新匹配的目标表行。使用 
      UPDATE SET column1 = source.column1 [, column2 = source.column2 ...]
      UPDATE SET column1 = source.column1 [, column2 = source.column2 ...]
      。
  • not_matched_action
    not_matched_action
    ： 
    • INSERT
      INSERT
      ：使用源数据集中的相应列插入目标表，支持指定字段插入。对于未指定的目标列，则插入为 
      NULL
      NULL
      。

注意事项

当执行 

MERGE INTO
MERGE INTO
语句可能产生不确定结果时，系统会报错。产生这种情况的原因有：

  1. 如果 
    ON
    ON
    子句使得源表中的多于 1 行与目标表中的行匹配，SQL 标准要求引发错误。
  2. 如果 
    ON
    ON
    子句使得源表中的多于 1 行与目标表中的行匹配，同时在 
    AND case_predicate
    AND case_predicate
    条件过滤后仍有多条记录。
  3. 当WHEN MATCHED语句和WHEN NOT MATCHED 语句同时存在时，MERGE INTO语句中只能存在一个WHEN NOT MATCHED
  4. 当存在多个 WHEN MATCHED 子句时，UPDATE 语句必须在 DELETE 语句之前。

确定性结果的情况：

  1. 如果 
    ON
    ON
    子句使得源表中的 1 行与目标表中的行匹配。
  2. 如果 
    ON
    ON
    子句使得源表中的多于 1 行与目标表中的行匹配，同时在 
    AND case_predicate
    AND case_predicate
    条件过滤后只有一条记录。

详细使用示例

案例1：当匹配条件满足时，删除目标表中的数据

场景：删除在源表中存在的所有用户记录（标记删除）

建表语句

-- 目标表：存储用户信息 CREATE TABLE users_target ( id INT, name STRING, value INT, updated_at STRING, marked_for_deletion BOOLEAN ); -- 源表：待同步的用户数据 CREATE TABLE users_source ( id INT, name STRING, value INT, updated_at STRING, marked_for_deletion BOOLEAN );

目标表插入的数据：

INSERT INTO users_target VALUES (1, 'Alice', 100, '2024-01-01', false), (2, 'Bob', 200, '2024-01-02', false), (3, 'Charlie', 300, '2024-01-03', false);

+----+---------+-------+------------+---------------------+ | id | name | value | updated_at | marked_for_deletion | +----+---------+-------+------------+---------------------+ | 1 | Alice | 100 | 2024-01-01 | false | +----+---------+-------+------------+---------------------+ | 2 | Bob | 200 | 2024-01-02 | false | +----+---------+-------+------------+---------------------+ | 3 | Charlie | 300 | 2024-01-03 | false | +----+---------+-------+------------+---------------------+

源表插入的数据：

INSERT INTO users_source VALUES (1, 'Alice', 100, '2024-01-01', false), (2, 'Bob', 200, '2024-01-02', false);

+----+-------+-------+------------+---------------------+ | id | name | value | updated_at | marked_for_deletion | +----+-------+-------+------------+---------------------+ | 1 | Alice | 100 | 2024-01-01 | false | +----+-------+-------+------------+---------------------+ | 2 | Bob | 200 | 2024-01-02 | false | +----+-------+-------+------------+---------------------+

MERGE SQL

MERGE INTO users_target t USING users_source s ON t.id = s.id WHEN MATCHED THEN DELETE;

执行结果

+----+---------+-------+------------+---------------------+ | id | name | value | updated_at | marked_for_deletion | +----+---------+-------+------------+---------------------+ | 3 | Charlie | 300 | 2024-01-03 | false | +----+---------+-------+------------+---------------------+

说明：源表中 id 为 1 和 2 的记录在目标表中找到匹配项，这两条记录被删除。id 为 3 的 Charlie 记录因为在源表中没有匹配项，保持不变。

案例2：基于时间戳的条件更新

场景：只在源表数据更新时间比目标表更新时间更新时，才更新目标表的值

建表语句

-- 目标表：产品信息 CREATE TABLE products_target ( id INT, name STRING, col1 STRING, updated_at STRING ); -- 源表：新的产品数据 CREATE TABLE products_source ( id INT, name STRING, col1 STRING, updated_at STRING );

初始目标表数据：

INSERT INTO products_target VALUES (1, 'Product A', 'value_old', '2024-01-01'), (2, 'Product B', 'value_old', '2024-01-02'), (3, 'Product C', 'value_old', '2024-01-03');

+----+------------+-----------+------------+ | id | name | col1 | updated_at | +----+------------+-----------+------------+ | 1 | Product A | value_old | 2024-01-01 | +----+------------+-----------+------------+ | 2 | Product B | value_old | 2024-01-02 | +----+------------+-----------+------------+ | 3 | Product C | value_old | 2024-01-03 | +----+------------+-----------+------------+

初始源表数据：

INSERT INTO products_source VALUES (1, 'Product A', 'value_new', '2024-01-10'), (2, 'Product B', 'value_old', '2024-01-01'), (4, 'Product D', 'value_new', '2024-01-04');

+----+------------+-----------+------------+ | id | name | col1 | updated_at | +----+------------+-----------+------------+ | 1 | Product A | value_new | 2024-01-10 | +----+------------+-----------+------------+ | 2 | Product B | value_old | 2024-01-01 | +----+------------+-----------+------------+ | 4 | Product D | value_new | 2024-01-04 | +----+------------+-----------+------------+

MERGE SQL

MERGE INTO products_target t USING products_source s ON t.id = s.id WHEN MATCHED AND t.updated_at < s.updated_at THEN UPDATE SET t.col1 = s.col1;

执行结果

+----+------------+-----------+------------+ | id | name | col1 | updated_at | +----+------------+-----------+------------+ | 1 | Product A | value_new | 2024-01-01 | +----+------------+-----------+------------+ | 2 | Product B | value_old | 2024-01-02 | +----+------------+-----------+------------+ | 3 | Product C | value_old | 2024-01-03 | +----+------------+-----------+------------+

说明：

  • Product A (id=1)：源表的更新时间 
    2024-01-10
    2024-01-10
    > 目标表的 
    2024-01-01
    2024-01-01
    ，条件满足，col1 被更新为 
    value_new
    value_new
  • Product B (id=2)：源表的更新时间 
    2024-01-01
    2024-01-01
    ≤ 目标表的 
    2024-01-02
    2024-01-02
    ，条件不满足，保持不变
  • Product C (id=3)：源表中没有对应记录（id=3），不进行任何操作，保持不变
  • Product D (id=4)：源表新增，由于不存在 ON 条件的 WHEN NOT MATCHED 子句，不插入

案例3：删除标记与条件更新的组合

场景：对用户数据进行操作，标记为删除的用户直接删除，其他匹配的用户更新其状态信息

建表语句

-- 目标表：用户信息 CREATE TABLE users_data_target ( id INT, name STRING, marked_for_deletion BOOLEAN, updated_at STRING, value STRING ); -- 源表：同步的用户数据 CREATE TABLE users_data_source ( id INT, name STRING, marked_for_deletion BOOLEAN, updated_at STRING, value STRING );

初始目标表数据：

INSERT INTO users_data_target VALUES (1, 'User A', false, '2024-01-01', 'active'), (2, 'User B', true, '2024-01-02', 'inactive'), (3, 'User C', false, '2024-01-03', 'active'), (4, 'User D', true, '2024-01-04', 'inactive');

+----+--------+---------------------+------------+----------+ | id | name | marked_for_deletion | updated_at | value | +----+--------+---------------------+------------+----------+ | 1 | User A | false | 2024-01-01 | active | +----+--------+---------------------+------------+----------+ | 2 | User B | true | 2024-01-02 | inactive | +----+--------+---------------------+------------+----------+ | 3 | User C | false | 2024-01-03 | active | +----+--------+---------------------+------------+----------+ | 4 | User D | true | 2024-01-04 | inactive | +----+--------+---------------------+------------+----------+

初始源表数据：

INSERT INTO users_data_source VALUES (1, 'User A', false, '2024-01-15', 'updated'), (2, 'User B', true, '2024-01-20', 'to_delete'), (3, 'User C', false, '2024-01-10', 'updated'), (5, 'User E', false, '2024-01-05', 'new_user');

+----+--------+---------------------+------------+-----------+ | id | name | marked_for_deletion | updated_at | value | +----+--------+---------------------+------------+-----------+ | 1 | User A | false | 2024-01-15 | updated | +----+--------+---------------------+------------+-----------+ | 2 | User B | true | 2024-01-20 | to_delete | +----+--------+---------------------+------------+-----------+ | 3 | User C | false | 2024-01-10 | updated | +----+--------+---------------------+------------+-----------+ | 5 | User E | false | 2024-01-05 | new_user | +----+--------+---------------------+------------+-----------+

MERGE SQL

MERGE INTO users_data_target t USING users_data_source s ON t.id = s.id WHEN MATCHED AND t.marked_for_deletion = false THEN UPDATE SET t.updated_at = s.updated_at, t.value = s.value WHEN MATCHED AND t.marked_for_deletion = true THEN DELETE;

执行结果

+----+--------+---------------------+------------+---------+ | id | name | marked_for_deletion | updated_at | value | +----+--------+---------------------+------------+---------+ | 1 | User A | false | 2024-01-15 | updated | +----+--------+---------------------+------------+---------+ | 3 | User C | false | 2024-01-10 | updated | +----+--------+---------------------+------------+---------+ | 4 | User D | true | 2024-01-04 | inactive| +----+--------+---------------------+------------+---------+

说明：

  • User A (id=1)：
    marked_for_deletion=false
    marked_for_deletion=false
    ，满足第一个 WHEN MATCHED 条件，更新 
    updated_at
    updated_at
    和 
    value
    value
  • User B (id=2)：
    marked_for_deletion=true
    marked_for_deletion=true
    ，满足第二个 WHEN MATCHED 条件，被删除
  • User C (id=3)：
    marked_for_deletion=false
    marked_for_deletion=false
    ，满足第一个 WHEN MATCHED 条件，更新 
    updated_at
    updated_at
    和 
    value
    value
  • User D (id=4)：源表中没有对应记录，保持不变
  • User E (id=5)：源表新增，但没有 WHEN NOT MATCHED 子句，不插入

案例4：插入与更新的完整同步

场景：完整的数据同步操作，既要更新匹配的记录，也要插入新的记录

建表语句

-- 目标表：数据记录 CREATE TABLE data_sync_target ( id INT, col1 STRING, col2 STRING ); -- 源表：新数据 CREATE TABLE data_sync_source ( id INT, col1 STRING, col2 STRING );

初始目标表数据：

INSERT INTO data_sync_target VALUES (1, 'data_a', 'value_a'), (2, 'data_b', 'value_b'), (3, 'data_c', 'value_c');

+----+--------+---------+ | id | col1 | col2 | +----+--------+---------+ | 1 | data_a | value_a | +----+--------+---------+ | 2 | data_b | value_b | +----+--------+---------+ | 3 | data_c | value_c | +----+--------+---------+

初始源表数据：

INSERT INTO data_sync_source VALUES (1, 'updated_a', 'updated_value_a'), (2, 'updated_b', 'updated_value_b'), (4, 'new_data', 'new_value'), (5, 'new_data2', 'new_value2');

+----+----------+------------------+ | id | col1 | col2 | +----+----------+------------------+ | 1 | updated_a| updated_value_a | +----+----------+------------------+ | 2 | updated_b| updated_value_b | +----+----------+------------------+ | 4 | new_data | new_value | +----+----------+------------------+ | 5 | new_data2| new_value2 | +----+----------+------------------+

MERGE SQL

MERGE INTO data_sync_target t USING data_sync_source s ON t.id = s.id WHEN MATCHED THEN UPDATE SET t.col1 = s.col1, t.col2 = s.col2 WHEN NOT MATCHED THEN INSERT (id, col1, col2) VALUES (s.id, s.col1, s.col2);

执行结果

+----+----------+------------------+ | id | col1 | col2 | +----+----------+------------------+ | 1 | updated_a| updated_value_a | +----+----------+------------------+ | 2 | updated_b| updated_value_b | +----+----------+------------------+ | 3 | data_c | value_c | +----+----------+------------------+ | 4 | new_data | new_value | +----+----------+------------------+ | 5 | new_data2| new_value2 | +----+----------+------------------+

说明：

  • id=1, 2：源表中存在匹配记录，执行 UPDATE 操作，col1 和 col2 都被更新为源表的值
  • id=3：源表中没有对应记录，保持原值不变
  • id=4, 5：源表新增的记录，执行 INSERT 操作，插入到目标表中
DELETE
TRUNCATE
Yunqi © 2024 Yunqi, Inc. All Rights Reserved.

联系我们
预约咨询
微信咨询
电话咨询