count(1),count(*),count(rowid)

最新推荐文章于 2021-04-15 16:17:06 发布

weixin_34354173

最新推荐文章于 2021-04-15 16:17:06 发布

阅读量73

点赞数

CC 4.0 BY-SA版权

原文链接：http://www.cnblogs.com/zhanglt/archive/2013/05/04/3058925.html

本文通过在千万级数据表上进行三次不同的SQL计数操作测试(count(1), count(*), count(rowid))，对比了其执行效率。结果显示，count(rowid)的执行速度最快，其次是count(1)，最慢的是count(*)。

下面是三个语句的测试结果，测试表的数据为25341223行（千万级）

SQL> alter system flush shared_pool;

System altered

Executed in 0.047 seconds

SQL> select count(1) from thqlog;

COUNT(1)
----------
25341223

Executed in 12.297 seconds

SQL> alter system flush shared_pool;

System altered

Executed in 0.062 seconds

SQL> select count(*) from thqlog
2 ;

COUNT(*)
----------
25341223

Executed in 12.563 seconds

SQL> alter system flush shared_pool;

System altered

Executed in 0.047 seconds

SQL> select count(rowid) from thqlog;

COUNT(ROWID)
------------
25341223

Executed in 12.172 seconds