[oracle/sql]关于清除重复，not in方案和not exists方案的对比

有这样一张表：

CREATE TABLE tb_sc
(
    id NUMBER not null primary key,
    studentid int not null,
    courseid int not null,
    score int not null
)

用以下语句给它充值十万条数据：

Insert into tb_sc
select rownum,dbms_random.value(0,10000),dbms_random.value(1,5),dbms_random.value(0,150) from dual
connect by level<=100000
order by dbms_random.random

当然上面这样填充完会给同一studentid和courseid时不同的score记录，相当于学生同一科目考了两次，这在现实中是不合理的，因此我们要剔除掉，只保留同一studentid和courseid时score最高的那条记录。

用下面的sql能查询出该保留的记录：

select studentid,courseid,max(score) as score from tb_sc group by studentid,courseid

如果直接去删除不在上面的sql查出的结果集中的记录，会写出如下的sql：

delete from tb_sc where (studentid,courseid,score) not in (select studentid,courseid,max(score) as score from tb_sc group by studentid,courseid)

但是，真执行起来就会发现，要等这条delete语句执行完简直遥遥无期。

而用同为反连接的not exist做就很快：

delete from tb_sc where not exists (
select 'x' from tb_sc a,
                 (select studentid,courseid,max(score) as score from tb_sc group by studentid,courseid) b
where a.studentid=b.studentid and a.courseid=b.courseid and a.score=b.score and tb_sc.id=a.id)

SQL> delete from tb_sc where not exists (
  2  select 'x' from tb_sc a,
  3                   (select studentid,courseid,max(score) as score from tb_sc group by studentid,courseid) b
  4  where a.studentid=b.studentid and a.courseid=b.courseid and a.score=b.score and tb_sc.id=a.id);

已删除58032行。

已用时间:  00: 00: 00.75

到这里很多人可能直接否决的not in，直接认为它慢，放弃了，但是看看下面sql，它会跑多久呢？

delete from tb_sc where id not in (select tb_sc.id from tb_sc,( select studentid,courseid,max(score) as score from tb_sc group by studentid,courseid ) tb_sc2
where tb_sc.studentid=tb_sc2.studentid and tb_sc.courseid=tb_sc2.courseid and tb_sc.score=tb_sc2.score)

让我们实际跑一下：

SQL> delete from tb_sc where id not in (select tb_sc.id from tb_sc,( select studentid,courseid,max(score) as score from tb_sc group by studentid,courseid ) tb_sc2
  2  where tb_sc.studentid=tb_sc2.studentid and tb_sc.courseid=tb_sc2.courseid and tb_sc.score=tb_sc2.score);

已删除58032行。

已用时间:  00: 00: 00.56

发现和上面的not exist差不多！你是不是又恢复对not in的信心了呢？

但是

delete from tb_sc where (studentid,courseid,score) not in (select studentid,courseid,max(score) as score from tb_sc group by studentid,courseid)

确实是慢得让人发指，而将delete换成select之后，却并不慢，不信大家请执行下面sql：

select a.* from tb_sc a where (a.studentid,a.courseid,a.score) not in (select studentid,courseid,max(score) as score from tb_sc group by studentid,courseid)

我执行的结果是：

     19563       9998          2          6
     27799       9998          3         95

        ID  STUDENTID   COURSEID      SCORE
---------- ---------- ---------- ----------
     37515       9998          4         60
     35809       9998          4        144
     65663       9998          5         64
     84961       9999          3          2
     24730       9999          3         14
     99371       9999          3         16
      4349      10000          5          2

已选择58032行。

已用时间:  00: 00: 16.01

并没有多长时间，为什么同样的条件，查询就快，而删除就慢得不要不要的，其中的原因还有待继续研究。

--2020年1月24日--