导航：首页 > 数据库 >

数据库中plan_hash_value有什么用

发表于：2024-11-26 作者：千家信息网编辑

千家信息网最后更新 2024年11月26日，这篇文章给大家分享的是有关数据库中plan_hash_value有什么用的内容。小编觉得挺实用的，因此分享给大家做个参考，一起跟随小编过来看看吧。plan_hash_value判断sql的访问路径是否

千家信息网最后更新 2024年11月26日数据库中plan_hash_value有什么用

这篇文章给大家分享的是有关数据库中plan_hash_value有什么用的内容。小编觉得挺实用的，因此分享给大家做个参考，一起跟随小编过来看看吧。

plan_hash_value

判断sql的访问路径是否改变的主要方法是:v$sql.plan_hash_value的值是否改变。如果不同的sql语句含有相同的

实验：
---创建表dept
SQL> create table dept as select * from scott.dept;

Table created.

---执行2条sql查询dept表
SQL> select deptno,dname from dept where deptno=10;

DEPTNO DNAME
---------- --------------
10 ACCOUNTING

SQL> select deptno,dname from dept;

DEPTNO DNAME
---------- --------------
10 ACCOUNTING
20 RESEARCH
30 SALES
40 OPERATIONS

---通过v$sql查询关于dept的sql的address,hash_value,child_number,plan_hash_value,sql_text
col SQL_TEXT for a100
select address,hash_value,child_number,plan_hash_value,sql_text from v$sql where sql_text like '%DEPT%';

ADDRESS HASH_VALUE CHILD_NUMBER PLAN_HASH_VALUE SQL_TEXT
---------------- ---------- ------------ --------------- ----------------------------------------------------------------------------------------------------
000000008B589B40 3749466620 0 315352865 SELECT /* OPT_DYN_SAMP */ /*+ ALL_ROWS IGNORE_WHERE_CLAUSE NO_PARALLEL(SAMPLESUB) opt_param('paralle
l_execution_enabled', 'false') NO_PARALLEL_INDEX(SAMPLESUB) NO_SQL_TUNE */ NVL(SUM(C1),0), NVL(SUM(C
2),0) FROM (SELECT /*+ NO_PARALLEL("DEPT") FULL("DEPT") NO_PARALLEL_INDEX("DEPT") */ 1 AS C1, 1 AS C
2 FROM "SYS"."DEPT" "DEPT") SAMPLESUB

0000000061211A40 2958346034 0 2657262937 select ADDRESS, HASH_VALUE, SQL_ID, PLAN_HASH_VALUE, CHILD_ADDRESS, CHILD_NUMBER, TIMESTAM
P, OPERATION, OPTIONS, OBJECT_NODE, OBJECT#, OBJECT_OWNER, O
BJECT_NAME, OBJECT_ALIAS, OBJECT_TYPE, OPTIMIZER,
ID, PARENT_ID, DEPTH, POSITION, SEARCH_COLUMNS, COST, CARDINALITY, BYTES, OTHER_TA
G, PARTITION_START, PARTITION_STOP, PARTITION_ID, OTHER, DISTRIBUTION, CPU_COST, IO_COST
, TEMP_SPACE, ACCESS_PREDICATES, FILTER_PREDICATES, PROJECTION, TIME, QBLOC
K_NAME, REMARKS, OTHER_XML from GV
$SQL_PLAN where inst_id = USERENV('Insta
nce')

000000008B5D3908 3410315986 0 903671040 select address,hash_value,child_number,plan_hash_value,sql_text from v$sql where sql_text like '%DEP
T%'

000000008B626668 3145184715 0 315352865 SELECT /* OPT_DYN_SAMP */ /*+ ALL_ROWS IGNORE_WHERE_CLAUSE NO_PARALLEL(SAMPLESUB) opt_param('paralle
l_execution_enabled', 'false') NO_PARALLEL_INDEX(SAMPLESUB) NO_SQL_TUNE */ NVL(SUM(C1),0), NVL(SUM(C
2),0) FROM (SELECT /*+ IGNORE_WHERE_CLAUSE NO_PARALLEL("DEPT") FULL("DEPT") NO_PARALLEL_INDEX("DEPT"
) */ 1 AS C1, CASE WHEN "DEPT"."DEPTNO"=10 THEN 1 ELSE 0 END AS C2 FROM "SYS"."DEPT" "DEPT") SAMPLES
UB

---通过v$sql_plan查询执行计划
SQL> SELECT operation, options, object_name, cost FROM v$sql_plan WHERE address ='000000008B589B40' and HASH_VALUE='3749466620';

OPERATION OPTIONS OBJECT_NAME COST
------------------------------------------------------------ ------------------------------------------------------------
SELECT STATEMENT 2
SORT AGGREGATE
TABLE ACCESS FULL DEPT 2

SQL>
SQL> SELECT operation, options, object_name, cost FROM v$sql_plan WHERE address ='000000008B626668' and HASH_VALUE='3145184715';

OPERATION OPTIONS OBJECT_NAME COST
------------------------------------------------------------ ------------------------------------------------------------ ------------------------------ ----------
SELECT STATEMENT 2
SORT AGGREGATE
TABLE ACCESS FULL DEPT 2
查询的执行计划完全一样

--对表dept做修改，增加索引
SQL> create index idx_dept_deptno on dept(deptno);

Index created.

---再次查看
SQL> select address,hash_value,child_number,plan_hash_value,sql_text from v$sql where sql_text like '%DEPT%';

ADDRESS HASH_VALUE CHILD_NUMBER PLAN_HASH_VALUE SQL_TEXT
---------------- ---------- ------------ --------------- ----------------------------------------------------------------------------------------------------
000000008B589B40 3749466620 0 315352865 SELECT /* OPT_DYN_SAMP */ /*+ ALL_ROWS IGNORE_WHERE_CLAUSE NO_PARALLEL(SAMPLESUB) opt_param('paralle
l_execution_enabled', 'false') NO_PARALLEL_INDEX(SAMPLESUB) NO_SQL_TUNE */ NVL(SUM(C1),0), NVL(SUM(C
2),0) FROM (SELECT /*+ NO_PARALLEL("DEPT") FULL("DEPT") NO_PARALLEL_INDEX("DEPT") */ 1 AS C1, 1 AS C
2 FROM "SYS"."DEPT" "DEPT") SAMPLESUB

0000000061211A40 2958346034 0 2657262937 select ADDRESS, HASH_VALUE, SQL_ID, PLAN_HASH_VALUE, CHILD_ADDRESS, CHILD_NUMBER, TIMESTAM
P, OPERATION, OPTIONS, OBJECT_NODE, OBJECT#, OBJECT_OWNER, O
BJECT_NAME, OBJECT_ALIAS, OBJECT_TYPE, OPTIMIZER,
ID, PARENT_ID, DEPTH, POSITION, SEARCH_COLUMNS, COST, CARDINALITY, BYTES, OTHER_TA
G, PARTITION_START, PARTITION_STOP, PARTITION_ID, OTHER, DISTRIBUTION, CPU_COST, IO_COST
, TEMP_SPACE, ACCESS_PREDICATES, FILTER_PREDICATES, PROJECTION, TIME, QBLOC
K_NAME, REMARKS, OTHER_XML from GV
$SQL_PLAN where inst_id = USERENV('Insta
nce')

000000008B5D3908 3410315986 0 903671040 select address,hash_value,child_number,plan_hash_value,sql_text from v$sql where sql_text like '%DEP
T%'

000000008E0A74E0 1470990285 0 0 LOCK TABLE "DEPT" IN SHARE MODE NOWAIT
000000008B626668 3145184715 0 315352865 SELECT /* OPT_DYN_SAMP */ /*+ ALL_ROWS IGNORE_WHERE_CLAUSE NO_PARALLEL(SAMPLESUB) opt_param('paralle
l_execution_enabled', 'false') NO_PARALLEL_INDEX(SAMPLESUB) NO_SQL_TUNE */ NVL(SUM(C1),0), NVL(SUM(C
2),0) FROM (SELECT /*+ IGNORE_WHERE_CLAUSE NO_PARALLEL("DEPT") FULL("DEPT") NO_PARALLEL_INDEX("DEPT"
) */ 1 AS C1, CASE WHEN "DEPT"."DEPTNO"=10 THEN 1 ELSE 0 END AS C2 FROM "SYS"."DEPT" "DEPT") SAMPLES
UB

----再次执行上述相同的2条sql
SQL> select deptno,dname from dept where deptno=10;

DEPTNO DNAME
---------- --------------
10 ACCOUNTING

SQL> select deptno,dname from dept;

DEPTNO DNAME
---------- --------------
10 ACCOUNTING
20 RESEARCH
30 SALES
40 OPERATIONS

---查询v$sql下的address,hash_value,child_number,plan_hash_value,sql_text
SQL> select address,hash_value,child_number,plan_hash_value,sql_text from v$sql where sql_text like '%DEPT%';

ADDRESS HASH_VALUE CHILD_NUMBER PLAN_HASH_VALUE SQL_TEXT
---------------- ---------- ------------ --------------- ----------------------------------------------------------------------------------------------------
000000008B589B40 3749466620 0 315352865 SELECT /* OPT_DYN_SAMP */ /*+ ALL_ROWS IGNORE_WHERE_CLAUSE NO_PARALLEL(SAMPLESUB) opt_param('paralle
l_execution_enabled', 'false') NO_PARALLEL_INDEX(SAMPLESUB) NO_SQL_TUNE */ NVL(SUM(C1),0), NVL(SUM(C
2),0) FROM (SELECT /*+ NO_PARALLEL("DEPT") FULL("DEPT") NO_PARALLEL_INDEX("DEPT") */ 1 AS C1, 1 AS C
2 FROM "SYS"."DEPT" "DEPT") SAMPLESUB

0000000061211A40 2958346034 0 2657262937 select ADDRESS, HASH_VALUE, SQL_ID, PLAN_HASH_VALUE, CHILD_ADDRESS, CHILD_NUMBER, TIMESTAM
P, OPERATION, OPTIONS, OBJECT_NODE, OBJECT#, OBJECT_OWNER, O
BJECT_NAME, OBJECT_ALIAS, OBJECT_TYPE, OPTIMIZER,
ID, PARENT_ID, DEPTH, POSITION, SEARCH_COLUMNS, COST, CARDINALITY, BYTES, OTHER_TA
G, PARTITION_START, PARTITION_STOP, PARTITION_ID, OTHER, DISTRIBUTION, CPU_COST, IO_COST
, TEMP_SPACE, ACCESS_PREDICATES, FILTER_PREDICATES, PROJECTION, TIME, QBLOC
K_NAME, REMARKS, OTHER_XML from GV
$SQL_PLAN where inst_id = USERENV('Insta
nce')

000000008B5D3908 3410315986 0 903671040 select address,hash_value,child_number,plan_hash_value,sql_text from v$sql where sql_text like '%DEP
T%'

000000008E0A74E0 1470990285 0 0 LOCK TABLE "DEPT" IN SHARE MODE NOWAIT
000000008B626668 3145184715 0 315352865 SELECT /* OPT_DYN_SAMP */ /*+ ALL_ROWS IGNORE_WHERE_CLAUSE NO_PARALLEL(SAMPLESUB) opt_param('paralle
l_execution_enabled', 'false') NO_PARALLEL_INDEX(SAMPLESUB) NO_SQL_TUNE */ NVL(SUM(C1),0), NVL(SUM(C
2),0) FROM (SELECT /*+ IGNORE_WHERE_CLAUSE NO_PARALLEL("DEPT") FULL("DEPT") NO_PARALLEL_INDEX("DEPT"
) */ 1 AS C1, CASE WHEN "DEPT"."DEPTNO"=10 THEN 1 ELSE 0 END AS C2 FROM "SYS"."DEPT" "DEPT") SAMPLES
UB

00000000613ACE30 1756886759 0 2570986044 SELECT /* OPT_DYN_SAMP */ /*+ ALL_ROWS opt_param('parallel_execution_enabled', 'false') NO_PARALLEL(
SAMPLESUB) NO_PARALLEL_INDEX(SAMPLESUB) NO_SQL_TUNE */ NVL(SUM(C1),0), NVL(SUM(C2),0), NVL(SUM(C3),0
) FROM (SELECT /*+ NO_PARALLEL("DEPT") INDEX("DEPT" IDX_DEPT_DEPTNO) NO_PARALLEL_INDEX("DEPT") */ 1
AS C1, 1 AS C2, 1 AS C3 FROM "SYS"."DEPT" "DEPT" WHERE "DEPT"."DEPTNO"=10 AND ROWNUM <= 2500) SAMPL
ESUB

6 rows selected.

---再次查看执行计划
SQL> SELECT operation, options, object_name, cost FROM v$sql_plan WHERE address ='000000008B589B40' and HASH_VALUE='3749466620';

OPERATION OPTIONS OBJECT_NAME COST
------------------------------------------------------------ ------------------------------------------------------------ ------------------------------ ----------
SELECT STATEMENT 2
SORT AGGREGATE
TABLE ACCESS FULL DEPT 2

SQL> SELECT operation, options, object_name, cost FROM v$sql_plan WHERE address ='000000008B626668' and HASH_VALUE='3145184715';

OPERATION OPTIONS OBJECT_NAME COST
------------------------------------------------------------ ------------------------------------------------------------ ------------------------------ ----------
SELECT STATEMENT 2
SORT AGGREGATE
TABLE ACCESS FULL DEPT 2

SQL> SELECT operation, options, object_name, cost FROM v$sql_plan WHERE address ='00000000613ACE30' and HASH_VALUE='1756886759';

OPERATION OPTIONS OBJECT_NAME COST
------------------------------------------------------------ ------------------------------------------------------------ ------------------------------ ----------
SELECT STATEMENT 1
SORT AGGREGATE
VIEW 1
COUNT STOPKEY
INDEX RANGE SCAN IDX_DEPT_DEPTNO 1
最终发现执行计划已经改变，所以可以根据plan_hash_value值是否变化可知该sql语句的执行计划是否改变

通过v$sql视图的plan_hash_value值可以很方便的知道，该sql语句的执行计划是否改变，
通常我们的做法是分别将v$sql,v$sql_hash做两份快照，隔段时间后再做段快照，然后将其进行比较，查看是否有sql的执行计划已经改变。