Question

我有这种工作的DB2请求

MERGE INTO table_to_upsert AS tab
USING (VALUES
        (1, 2, 3),
        (4, 5, 6),
        (7, 8, 9)
        -- more rows
    ) AS merge (C1, C2, C3)
    ON tab.key_to_match = merge.key_to_match -- usually the primary key example : tab.C2 = merge.C2
    WHEN MATCHED THEN
        UPDATE SET tab.C1 = merge.C1,
                   tab.C2 = merge.C2,
                   tab.C3 = merge.C3
    WHEN NOT MATCHED THEN
        INSERT (C1, C2, C3)
        VALUES (merge.C1, merge.C2, merge.C3)

但是现在，我想要的是，IF主键已经存在，然后更新，但只有在C1的新值与旧值不同时才更新C1。

在理论上类似的东西

MERGE INTO table_to_upsert AS tab
USING (VALUES
        (1, 2, 3),
        (4, 5, 6),
        (7, 8, 9)
        -- more rows
    ) AS merge (C1, C2, C3)
    ON tab.key_to_match = merge.key_to_match
    WHEN MATCHED THEN
        UPDATE SET tab.C1 = merge.C1,
                   tab.C2 = merge.C2,
                   tab.C3 = merge.C3
        WHERE tab.C1 != merge.C1
    WHEN NOT MATCHED THEN
        INSERT (C1, C2, C3)
        VALUES (merge.C1, merge.C2, merge.C3)

我添加了WHERE tab.C1 != merge.C1，但它无效，导致

错误：ILLEGAL SYMBOL WHERE;有效符号

Answer 1

如果我没记错的话，你可以在匹配的行中添加一个搜索条件，即

MERGE INTO table_to_upsert AS tab
USING (VALUES
        (1, 2, 3),
        (4, 5, 6),
        (7, 8, 9)
        -- more rows
    ) AS merge (C1, C2, C3)
    ON tab.key_to_match = merge.key_to_match
    WHEN MATCHED AND tab.C1 != merge.C1 THEN
        UPDATE SET tab.C1 = merge.C1,
                   tab.C2 = merge.C2,
                   tab.C3 = merge.C3
    WHEN NOT MATCHED THEN
        INSERT (C1, C2, C3)
        VALUES (merge.C1, merge.C2, merge.C3)

DB2，MERGE INTO如果值不同则更新

1 个答案: