5

我想在 MERGE 语句中更新多个列,但是对于每一列,条件都不同。我该如何实现。

我在两个表中都有超过 100 万行。所有列都是数字。除了 Id,所有 3 列的数字都精度约为 18 位,例如:1.34255353433230675

有没有更好的更新方法。每天可能会更新大约 50,000 行所以我必须将更新值合并到目标表。我尝试了 UPDATE 和 FORALL,但速度很慢。

我基本上想根据公共 ID 列合并差异。还有其他方法更好吗?

DECLARE
TYPE test1_t IS TABLE OF test.score%TYPE INDEX BY PLS_INTEGER;
TYPE test2_t IS TABLE OF test.id%TYPE INDEX BY PLS_INTEGER;
TYPE test3_t IS TABLE OF test.Crank%TYPE INDEX BY PLS_INTEGER;
TYPE test4_t IS TABLE OF test.urank%TYPE INDEX BY PLS_INTEGER;

vscore    test1_t;
vid       test2_t;
vcrank    test3_t;
vurank    test4_t;
BEGIN
 SELECT id,score,crank,urank
 BULK   COLLECT INTO vid,vscore,vcrank,vurank
 FROM   test;

 FORALL i IN 1 .. vid.COUNT
  MERGE INTO final T
  USING      (SELECT vid (i) AS o_id,
                     vcrank (i) AS o_crank,
                     vurank (i) AS o_urank
                     vscore (i) AS o_score
              FROM   DUAL) S
  ON         (S.o_id = T.id)
  WHEN MATCHED THEN
  UPDATE SET T.crank = S.o_crank
  WHERE  T.crank <> S.o_crank;

  UPDATE SET T.crank = S.o_crank
  WHERE  T.crank <> S.o_crank;

  UPDATE SET T.crank = S.o_crank
  WHERE  T.crank <> S.o_crank;

  UPDATE SET T.score = S.score
  WHERE  T.score <> S.score;

 -- I tried the below case its not working either...
--   UPDATE SET T.crank = (CASE WHEN T.crank <> S.o_crank
--                        THEN S.o_crank
--                        END),
--           T.urank = (CASE WHEN T.urank <> S.o_urank
--                        THEN S.o_urank
--                        END);  

 COMMIT;
END;

/

4

1 回答 1

8

我认为你不需要循环。我假设您的 id 是主键,并且您并不是要在示例中重复多次曲柄。

像这样的东西会起作用吗?

根据 Raj A 的评论进行编辑。这只会更新其他字段之一已更改的行。请注意,这不会更新其中一个为 NULL 而另一个不为 NULL 的行。

MERGE INTO final T 
USING ( SELECT id, score, crank, urank FROM test ) S
   ON ( S.vid = T.id AND 
        ( S.crank != T.crank OR S.score != T.score OR S.urank != T.urank ))
 WHEN MATCHED SET crank = S.crank, score = S.score, 
      crank = S.crank, urank = S.urank 
 WHEN NOT MATCHED THEN INSERT
      [... not sure what you want to do in this case ...]
于 2013-06-15T02:16:35.757 回答