SQL> desc v$sqltext;
Name Null? Type
ADDRESS
HASH_VALUE
SQL_ID
COMMAND_TYPE NUMBER
PIECE NUMBER
SQL_TEXT
但v$sqltext中存储的比较简单，没有该语句的统计信息，比如执行次数等。其中piece表示SQL语句分片之后的顺序编号，比如三行的值为0，1，2, 按照顺序连接起来就是一个完成的SQL语句。sql_text 表示分片后的sql语句的一部分，注意它的长度只有64 bytes . 字段HASH_VALUE 和 address 一起唯一标志一条sql 。
下面的sql，利用v$session视图里的 PREV_SQL_ADDR， PREV_HASH_VALUE列，确定某个SID或者spid对应的完整SQL
（1）根据sid获得sql语句
select /*+ ORDERED */ sql_text
from v$sqltext
where (a.hash_value,a.ADDRESS) in
( select decode(sql_hash_value,0,PREV_HASH_VALUE,sql_hash_value),
decode(sql_hash_value,0,PREV_SQL_ADDR,SQL_ADDRESS)
from v$session
where b.sid=&SID)
order by piece asc
SQL> select sid,username from v$session;
SID USERNAME
------- --------
36
37 SYS --要查询的当前SYS用户在执行什么语句
38
40
26 rows selected.
SQL> select /*+ ORDERED */ sql_text
from v$sqltext a
where (a.hash_value,a.ADDRESS) in
( select decode(sql_hash_value,0,PREV_HASH_VALUE,sql_hash_value),
decode(sql_hash_value,0,PREV_SQL_ADDR,SQL_ADDRESS)
from v$session b
where b.sid=&SID)
order by piece asc;
Enter value for sid: 37
old 7: where b.sid=&SID)
new 7: where b.sid=37)
SQL_TEXT
---------
select /*+ ORDERED */ sql_text from v$sqltext a where (a.hash_va
lue,a.ADDRESS) in ( select decode(sql_hash_value,0,PREV_HASH_VAL
UE,sql_hash_value), decode(sql_hash_value,0,PREV_SQL_ADDR,SQL_AD
DRESS) from v$session b where b.sid=37) order by piece asc
（2）根据操作系统pid获得sql语句，这个进程是oracle进程并且是LOCAL=NO
select /*+ ORDERED */ sql_text
from v$sqltext a
where (a.hash_value,a.ADDRESS) in
( select decode(sql_hash_value,0,PREV_HASH_VALUE,sql_hash_value),
decode(sql_hash_value,0,PREV_SQL_ADDR,SQL_ADDRESS)
from v$session b
where b.paddr = (select addr
from v$process c
where c.spid = '&spid'))
order by piece asc
以操作系统的process ID为参数。（这个可以通过top命令找出LOCAL=NO的oracle消耗CPU较高的进程PID）
下面是实际生产环境的例子，业务反馈CPU使用率过高，登入到主机使用TOP查看，可以看到使用top命令查看到PID为3193,31433,31690这几个oracle进程使用CPU达到了百分之百了，这里可以使用上面的v$sqltext视图结合v$session和v$process来定位正在运行的占用cpu较高的SQL语句。
oracle@ZJHZ-HW-ZQRZ-WSOP-2:~> top
top - 15:50:18 up 1041 days, 23:28, 6 users, load average: 11.93, 8.62, 14.23
Tasks: 509 total, 18 running, 490 sleeping, 0 stopped, 1 zombie
Cpu(s): 89.2%us, 1.2%sy, 0.2%ni, 9.1%id, 0.4%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 23641M total, 22667M used, 973M free, 1534M buffers
Swap: 25583M total, 80M used, 25503M free, 17946M cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
31493 oracle 20 0 16.3g 245m 241m R 100 1.0 3:18.24 oracle
31433 oracle 20 0 16.3g 246m 242m R 100 1.0 3:24.92 oracle
31690 oracle 20 0 16.3g 250m 246m R 100 1.1 3:11.16 oracle
31609 oracle 20 0 16.3g 245m 241m R 99 1.0 3:15.02 oracle
9645 oracle 20 0 16.3g 320m 316m R 99 1.4 3:33.72 oracle
11860 oracle 20 0 16.3g 374m 370m R 98 1.6 3:44.13 oracle
31553 oracle 20 0 16.3g 243m 239m R 98 1.0 3:22.38 oracle
31222 oracle 20 0 16.3g 247m 243m R 97 1.0 3:25.97 oracle
456 oracle 20 0 16.3g 321m 316m R 97 1.4 3:36.43 oracle
31127 oracle 20 0 16.3g 247m 243m R 95 1.0 3:29.34 oracle
5205 oracle 20 0 2958m 79m 18m R 75 0.3 0:02.25 java
5173 oracle 20 0 98684 21m 7896 S 6 0.1 0:00.18 perl
11357 root 20 0 672m 9556 6232 S 3 0.0 27:11.31 InforGuardMa
12016 root 30 10 470m 8256 3212 S 3 0.0 4909:32 hpi_program
26379 root 20 0 3876 468 372 S 1 0.0 4711:34 guard-userspace
5162 oracle 20 0 9072 1532 872 R 1 0.0 0:00.03 top
11425 root 20 0 223m 60m 6440 S 1 0.3 1299:53 java
19249 oracle -2 0 16.3g 15m 14m S 1 0.1 1473:10 oracle
30529 oracle 20 0 9072 1544 876 S 1 0.0 0:01.42 top
oracle@ZJHZ-HW-ZQRZ-WSOP-2:~> sqlplus / as sysdba
SQL*Plus: Release 11.2.0.3.0 Production on Tue Aug 28 15:50:34 2018
Copyright (c) 1982, 2011, Oracle. All rights reserved.
Connected to:
Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production
With the Partitioning, Automatic Storage Management, Oracle Label Security, OLAP,
Data Mining, Oracle Database Vault and Real Application Testing options
SQL> select /*+ ORDERED */ sql_text
from v$sqltext a
where (a.hash_value,a.ADDRESS) in
( select decode(sql_hash_value,0,PREV_HASH_VALUE,sql_hash_value),
decode(sql_hash_value,0,PREV_SQL_ADDR,SQL_ADDRESS)
from v$session b
where b.paddr = (select addr
from v$process c
where c.spid = '&spid'))
order by piece asc 2 3 4 5 6 7 8 9 10
11 ;
Enter value for spid: 31493 ---输入pid找出对应oracle进程执行sql语句
old 9: where c.spid = '&spid'))
new 9: where c.spid = '31493'))
SQL_TEXT ---这里定位出消耗CPU较高的SQL语句了，就是这条SQL语句导致CPU使用率达到百分之百，后面就要对这条SQL语句进行分析优化了。
----------------------------------------------------------------
select * from (select row_.*, rownum rownum_ from (
SELECT a.*, b.GROUPCUSTOMERNAME,b.GROUP
CUSTOMERID,m.MEMBERGROUPID FROM AAA_WSOPGROUPSUBSREG
ISTERLOG a JOIN AAA_WSOPGROUPCUSTOMER b ON a.GRO
UPCUSTOMERKEY=b.GROUPCUSTOMERKEY LEFT JOIN AAA_WSOPMEMBERGROU
P m ON a.MEMBERGROUPKEY=m.MEMBERGROUPKEY WHERE EXISTS (
SELECT :"SYS_B_0" FROM ( SELECT c.GROUP
CUSTOMERKEY FROM AAA_WSOPCUSTOMERORGREL c
WHERE EXISTS ( SELECT :"SYS_B_1" FROM (
SELECT ORGID, PARENTID
FROM T_BME_ORGANIZATION START WITH ORGI
SQL_TEXT
----------------------------------------------------------------
D = :1 CONNECT BY PRIOR ORGID = PARENTID
) d WHERE c.ORGID = d.ORGID
) ) e WHERE a.REGISTERTYPE != :"SYS_
B_2" AND a.STATUS != :"SYS_B_3" AND a.GROUPCUSTOMERKEY = e
.GROUPCUSTOMERKEY
) order by a.GROUPCUSTOMER
KEY,a.USERNAME )row_ where rownum <= :"SYS_B_4" ) whe
re rownum_ >= :"SYS_B_5"
20 rows selected.
上面例子还结合v$process这张视图，请参考我v$process这篇博客：Oracle v$PROCESS
另外v$sqltext诊断事件案例请参考我的博客： Oracle v$SQLTEXT案例
