count数据库优化oracle,对比各种count语句的效率和性能
欢迎大家仍砖头
[php]
测试环境windows xp + Oracle 9.2.0.1
在日常工作中，大家经常使用count.通常有以下几种写法：
count(*)
count(1)
count(rowid)
count('X')
count('column_name')
count('PK') --使用索引
这时候大家难免会争论一下哪个效率更高一点，有的人认为count(rowid)的效率更高，也有人持有不同意见。
那么究竟哪个效率更高，下面通过一个简单的实验来示范一下。
备注：请注意测试时使用的Oracle版本，在8以前稍有不同
Connected to Oracle9i Enterprise Edition Release 9.2.0.1.0
Connected as study
SQL> create table toms as select * from tomsrep.subs@from_ores1 where region=534 and rownum<100001;
Table created
SQL> alter table toms add constraint pk_toms primary key (oid);
Table altered
分别执行以下几个count语句，注意期间要清除data buffer，以保证数据环境的一致性！
SQL> select count(*) from toms;
count(*)
----------
100000
SQL> select count(1) from toms;
count(1)
----------
100000
SQL> select count('X') from toms;
count('X')
----------
100000
SQL> select count(rowid) from toms;
count(ROWID)
------------
100000
SQL> select count(oid) from toms;
count(OID)
----------
100000
SQL> select /*+ index(toms PK_TOMS) */ count(*) from toms;
COUNT(*)
----------
100000
对比每个count语句的统计信息如下：
select count(*) from toms
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.14 0.23 1 1 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.14 1.02 3948 3957 0 1
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.28 1.25 3949 3958 0 1
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 59 (STUDY)
Rows Row Source Operation
------- ---------------------------------------------------
1 SORT AGGREGATE
100000 TABLE ACCESS FULL TOMS
select count(1) from toms
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.01 0.02 1 1 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.12 1.08 3948 3957 0 1
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.14 1.10 3949 3958 0 1
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 59 (STUDY)
Rows Row Source Operation
------- ---------------------------------------------------
1 SORT AGGREGATE
100000 TABLE ACCESS FULL TOMS
select count(rowid) from toms
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.01 0.00 1 1 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.09 1.03 3948 3957 0 1
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.10 1.04 3949 3958 0 1
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 59 (STUDY)
Rows Row Source Operation
------- ---------------------------------------------------
1 SORT AGGREGATE
100000 TABLE ACCESS FULL TOMS
select count('X') from toms
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.01 0.02 1 1 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.17 1.03 3948 3957 0 1
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.18 1.05 3949 3958 0 1
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 59 (STUDY)
Rows Row Source Operation
------- ---------------------------------------------------
1 SORT AGGREGATE
100000 TABLE ACCESS FULL TOMS
select count(oid) from toms
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.00 0.02 1 1 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.20 1.03 3948 3957 0 1
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.20 1.05 3949 3958 0 1
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 59 (STUDY)
Rows Row Source Operation
------- ---------------------------------------------------
1 SORT AGGREGATE
100000 TABLE ACCESS FULL TOMS
select /*+ index(toms PK_TOMS) */ count(*) from toms
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.20 0.24 1 1 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.04 0.18 266 266 0 1
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.25 0.43 267 267 0 1
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 59 (STUDY)
Rows Row Source Operation
------- ---------------------------------------------------
1 SORT AGGREGATE
100000 INDEX FULL SCAN PK_TOMS (object id 34808)
整理如下：
对比可以看出，前5个count SQL语句的disk + query 都为3949 + 3957，这就保证了这几个SQL的执行环境
是一致的，同时也说明在性能上并没有明显差异
在一个有主键的表上count是可以只访问索引不访问表的，可以看出disk+query比使用全表扫描小些
在CPU消耗和总的时间消耗上有点不同：
count(*)： cpu=0.28 elapsed=1.25
count(1)： cpu=0.14 elapsed=1.10
count(‘X’)： cpu=0.18 elapsed=1.05
count(rowid)： cpu=0.10 elapsed=1.04
count(pk)： cpu=0.20 elapsed=1.05
count(pk:use index): cpu=0.25 elapsed0.43
对比一下，可以看出在CPU和总的时间消耗上相差甚微，LIO都相同性能上并没有明显差异。所以,我们可以
大胆的说他们的性能几乎是完全一样的, 大胆的猜测 Oracle 在9I[8i上没测试]以后Oracle在内部把count(*)、
count(1)、count(rowid)、count(pk)都统一转化为count(*)来处理,如果有PK或在有非空约束的字段上存在索引,
count(*)一样可以只访问该索引段得到数据.
所以大家今后对这种五花八门的count语句，也就不要见怪了，或许是每个人习惯不同而已。
最后：
在Oracle9i 8i以前，上述测试的结果可能不太相同，所以大家要永远记住一老话：实践是检验真理的唯一标准
[/php]