MySQL中对于not in和minus使用的优化

时间:2022-11-04 21:58:20

优化前:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
select count(t.id)
 from test t
 where t.status = 1
  and t.id not in (select distinct a.app_id
           from test2 a
           where a.type = 1
            and a.rule_id in (152, 153, 154))
      
 17:20:57 laojiu>@plan
 
PLAN_TABLE_OUTPUT
————————————————————————————————————————-
Plan hash value: 684502086
 
—————————————————————————————-
| Id | Operation      | Name       | Rows | Bytes | Cost (%CPU)| Time   |
—————————————————————————————-
|  0 | SELECT STATEMENT  |         |   1 |  18 |  176K (2)| 00:35:23 |
|  1 | SORT AGGREGATE   |         |   1 |  18 |      |     |
|* 2 |  FILTER      |         |    |    |      |     |
|* 3 |  TABLE ACCESS FULL| test   | 1141 | 20538 |  845  (2)| 00:00:11 |
|* 4 |  TABLE ACCESS FULL| test2 |   1 |  12 |  309  (2)| 00:00:04 |
—————————————————————————————-
 
Predicate Information (identified by operation id):
—————————————————
 
  2 – filter( NOT EXISTS (SELECT /*+ */ 0 FROM “test2″ “A” WHERE
       “A”.”type”=1 AND (“A”.”RULE_ID”=152 OR “A”.”RULE_ID”=153 OR
       “A”.”RULE_ID”=154) AND LNNVL(“A”.”APP_ID”<>:B1)))
  3 – filter(“T”.”status”=1)
  4 – filter(“A”.”type”=1 AND (“A”.”RULE_ID”=152 OR “A”.”RULE_ID”=153 OR
       “A”.”RULE_ID”=154) AND LNNVL(“A”.”APP_ID”<>:B1))
Statistics
———————————————————-
     0 recursive calls
     0 db block gets
  1762169 consistent gets
     0 physical reads
     0 redo size
    519 bytes sent via SQL*Net to client
    492 bytes received via SQL*Net from client
     2 SQL*Net roundtrips to/from client
     0 sorts (memory)
     0 sorts (disk)
     1 rows processed
21 rows selected.

优化后:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
select count(*) from(
 select t.id
  from test t
 where t.status = 1
 minus
 select distinct a.app_id
  from test2 a
 where a.type = 1
  and a.rule_id in (152, 153, 154))
17:23:33 laojiu>@plan
 
PLAN_TABLE_OUTPUT
————————————————————————————————————————-
Plan hash value: 631655686
 
————————————————————————————————–
| Id | Operation       | Name       | Rows | Bytes |TempSpc| Cost (%CPU)| Time   |
————————————————————————————————–
|  0 | SELECT STATEMENT   |         |   1 |    |    | 1501  (2)| 00:00:19 |
|  1 | SORT AGGREGATE    |         |   1 |    |    |      |     |
|  2 |  VIEW        |         | 1141 |    |    | 1501  (2)| 00:00:19 |
|  3 |  MINUS       |         |    |    |    |      |     |
|  4 |   SORT UNIQUE    |         | 1141 | 20538 |    |  846  (2)| 00:00:11 |
|* 5 |   TABLE ACCESS FULL| test   | 1141 | 20538 |    |  845  (2)| 00:00:11 |
|  6 |   SORT UNIQUE    |         | 69527 |  814K| 3632K|  654  (2)| 00:00:08 |
|* 7 |   TABLE ACCESS FULL| test2 | 84140 |  986K|    |  308  (2)| 00:00:04 |
————————————————————————————————–
 
Predicate Information (identified by operation id):
—————————————————
 
  5 – filter(“T”.”status”=1)
  7 – filter(“A”.”type”=1 AND (“A”.”RULE_ID”=152 OR “A”.”RULE_ID”=153 OR
       “A”.”RULE_ID”=154))
 
21 rows selected.
Statistics
———————————————————-
     1 recursive calls
     0 db block gets
    2240 consistent gets
     0 physical reads
     0 redo size
    516 bytes sent via SQL*Net to client
    492 bytes received via SQL*Net from client
     2 SQL*Net roundtrips to/from client
     2 sorts (memory)
     0 sorts (disk)
     1 rows processed

在优化sql的时候,我们需要转变一下思路,等价的改写sql;

改写后的sql由于逻辑读得到了天翻地覆的改变,很快得到结果。

第一条sql执行计划中有一个函数,LNNVL(“A”.”APP_ID”<>:B1),lnnvl(exp)

如果exp的结果是false或者是unknown,那么lnnvl返回true;

如果exp的结果是true,返回false.