Oracle sql 性能优化调整（14/15/16/17/18/19）

 14.       用Where子句替换HAVING子句

      避免使用HAVING子句, HAVING 只会在检索出所有记录之后才对结果集进行过滤. 这个处理需要排序,总计等操作. 如果能通过WHERE子句限制记录的数目,那就能减少这方面的开销.

 例如:

     低效:

     SELECT REGION，AVG(LOG_SIZE)

     FROM LOCATION

     GROUP BY REGION

     HAVING REGION REGION != ‘SYDNEY’

     AND REGION != ‘PERTH’

      高效

     SELECT REGION，AVG(LOG_SIZE)

     FROM LOCATION

     WHERE REGION REGION != ‘SYDNEY’

     AND REGION != ‘PERTH’

     GROUP BY REGION

( HAVING 中的条件一般用于对一些集合函数的比较,如COUNT() 等等. 除此而外,一般的条件应该写在WHERE子句中)

15.       减少对表的查询

在含有子查询的SQL语句中,要特别注意减少对表的查询.

  例如:

     低效

          SELECT TAB_NAME

          FROM TABLES

          WHERE TAB_NAME = ( SELECT TAB_NAME

                                FROM TAB_COLUMNS

                                WHERE VERSION = 604)

          ANDDB_VER= ( SELECT DB_VER

                           FROM TAB_COLUMNS

                           WHERE VERSION = 604)

     高效

          SELECT TAB_NAME

          FROM TABLES

          WHERE  (TAB_NAME,DB_VER)

 = ( SELECT TAB_NAME,DB_VER)

                   FROM TAB_COLUMNS

                   WHERE VERSION = 604)

     Update 多个Column 例子:

     低效:

           UPDATE EMP

           SET EMP_CAT = (SELECT MAX(CATEGORY) FROM EMP_CATEGORIES),

              SAL_RANGE = (SELECT MAX(SAL_RANGE) FROM EMP_CATEGORIES)

           WHERE EMP_DEPT = 0020;

     高效:

           UPDATE EMP

           SET (EMP_CAT, SAL_RANGE)

 = (SELECT MAX(CATEGORY) , MAX(SAL_RANGE)

 FROM EMP_CATEGORIES)

           WHERE EMP_DEPT = 0020;

   16.       通过内部函数提高SQL效率.

      SELECT H.EMPNO,E.ENAME,H.HIST_TYPE,T.TYPE_DESC,COUNT(*)

     FROM HISTORY_TYPE T,EMP E,EMP_HISTORY H

     WHERE H.EMPNO = E.EMPNO

AND H.HIST_TYPE = T.HIST_TYPE

GROUP BY H.EMPNO,E.ENAME,H.HIST_TYPE,T.TYPE_DESC;

通过调用下面的函数可以提高效率.

FUNCTION LOOKUP_HIST_TYPE(TYP IN NUMBER) RETURN VARCHAR2

AS

    TDESC VARCHAR2(30);

    CURSOR C1 IS  

        SELECT TYPE_DESC

        FROM HISTORY_TYPE

        WHERE HIST_TYPE = TYP;

BEGIN

    OPEN C1;

    FETCH C1 INTO TDESC;

    CLOSE C1;

    RETURN (NVL(TDESC,’?’));

END;

 

FUNCTION LOOKUP_EMP(EMP IN NUMBER) RETURN VARCHAR2

AS

    ENAME VARCHAR2(30);

    CURSOR C1 IS  

        SELECT ENAME

        FROM EMP

        WHERE EMPNO=EMP;

BEGIN

    OPEN C1;

    FETCH C1 INTO ENAME;

    CLOSE C1;

    RETURN (NVL(ENAME,’?’));

END;

 

SELECT H.EMPNO,LOOKUP_EMP(H.EMPNO),

H.HIST_TYPE,LOOKUP_HIST_TYPE(H.HIST_TYPE),COUNT(*)

FROM EMP_HISTORY H

GROUP BY H.EMPNO , H.HIST_TYPE;

17.       使用表的别名(Alias)

当在SQL语句中连接多个表时, 请使用表的别名并把别名前缀于每个Column上.这样一来,就可以减少解析的时间并减少那些由Column歧义引起的语法错误.

18.       用EXISTS替代IN

在许多基于基础表的查询中,为了满足一个条件,往往需要对另一个表进行联接.在这种情况下, 使用EXISTS(或NOT EXISTS)通常将提高查询的效率.

 低效:

SELECT *

FROM EMP (基础表)

WHERE EMPNO > 0

AND DEPTNO IN (SELECT DEPTNO

FROM DEPT

WHERE LOC = ‘MELB’)

    高效:

SELECT *

FROM EMP (基础表)

WHERE EMPNO > 0

AND EXISTS (SELECT ‘X’

FROM DEPT

WHERE DEPT.DEPTNO = EMP.DEPTNO

AND LOC = ‘MELB’)

( 相对来说,用NOT EXISTS替换NOT IN 将更显著地提高效率,下一节中将指出)

19.       用NOT EXISTS替代NOT IN

在子查询中,NOT IN子句将执行一个内部的排序和合并. 无论在哪种情况下,NOT IN都是最低效的 (因为它对子查询中的表执行了一个全表遍历).  为了避免使用NOT IN ,我们可以把它改写成外连接(Outer Joins)或NOT EXISTS.

 例如:

SELECT …

FROM EMP

WHERE DEPT_NO NOT IN (SELECT DEPT_NO

                         FROM DEPT

                         WHERE DEPT_CAT=’A’);

为了提高效率.改写为:

 (方法一: 高效)

SELECT ….

FROM EMP A,DEPT B

WHERE A.DEPT_NO = B.DEPT(+)

AND B.DEPT_NO IS NULL

AND B.DEPT_CAT(+) = ‘A’

 (方法二: 最高效)

SELECT ….

FROM EMP E

WHERE NOT EXISTS (SELECT ‘X’

                    FROM DEPT D

                    WHERE D.DEPT_NO = E.DEPT_NO

                    AND DEPT_CAT = ‘A’);