隐藏的方法调用代来性能问题。

简介: 原创 转载请注明出处 2010年6月13日 POMS生产故障分析及解决方法 一、问题描述和分析  2010年6月13日早上9点左右我们发现了生产库出现性能问题,随即我们查看其服务器性能发现CPU,I/O,内存利用率都基本正常。

原创 转载请注明出处

2010613 POMS生产故障分析及解决方法

一、问题描述和分析 

2010613早上9点左右我们发现了生产库出现性能问题,随即我们查看其服务器性能发现CPUI/O,内存利用率都基本正常。随后我们进行了初步分析,找到了几条问题SQL,但总结起来主要是如下两条:

1

SELECT COUNT(CPL.CHANGE_ID)
  FROM T_PRODUCT_FEE PF, T_POLICY_CHANGE PC, T_CONTRACT_PRODUCT_LOG CPL
 WHERE PF.POLICY_ID = PC.POLICY_ID
   AND PC.SERVICE_ID = 213
   AND PC.CHANGE_STATUS = 3
   AND PC.CHANGE_ID = CPL.CHANGE_ID
   AND PF.PRODUCT_NUM = CPL.PRODUCT_NUM
   AND PF.PRODUCT_ID = CPL.PRODUCT_ID
   AND PF.POLICY_ID = CPL.POLICY_ID
   AND PF.LIST_ID = :B1

2

SELECT DECODE(SUM(PF.PERIOD_PREM), NULL, 0, SUM(PF.PERIOD_PREM))

  FROM T_PRODUCT_FEE PF, T_CONTRACT_MASTER CM, T_PRODUCT_LIFE PL

 WHERE PF.POLICY_ID = CM.POLICY_ID

   AND CM.VALIDATE_DATE

   AND CM.VALIDATE_DATE >= ADD_MONTHS(TO_DATE(:B1, 'yyyymm'), -26)

   AND PF.CHARGE_TYPE = 1

   AND PF.PRODUCT_ID = PL.PRODUCT_ID

   AND (PL.PERIOD_TYPE 2 OR EXISTS

        (SELECT 1

           FROM T_LIFE_BASIC CC

          WHERE CC.COVERAGE_YEAR 1

            AND CC.PRODUCT_ID = PL.PRODUCT_ID))

   AND PF.FEE_TYPE IN (41, 47, 48)

   AND PF.POLICY_YEAR = 1

   AND PKG_LIFE_CALC_BASIC_COMM.F_IS_EFFECT(PF.LIST_ID) = '0'

   AND CM.PREM_STATUS NOT IN (2, 3, 4)

   AND CM.NB_AGENCY_ID IN

       (SELECT AGENCY_ID

          FROM T_AGENCY

         START WITH AGENCY_ID = :B2

        CONNECT BY PRIOR AGENCY_ID = PARENT_ID);

经过了和开发人员的仔细确认,确定这两个语句有一定的关联性,语句2正是通过方法

PKG_LIFE_CALC_BASIC_COMM.F_IS_EFFECT来调用的语句1,带入真实值后语句2的执行时间为15秒左右,如果多次调用势必会引起数据库性能问题,而事实就是如此,仔细分析语句2,语句2的执行计划如下:

 

Plan hash value: 1942344324

 

----------------------------------------------------------------------------------------------------

| Id  | Operation                         | Name                         | Starts | E-Rows | A-Rows |   A-Time   | Buffers |

----------------------------------------------------------------------------------------------------

|   1 |  SORT AGGREGATE                   |                              |      1 |      1 |      1 |00:00:09.91 |    1301K|   

|*  2 |   FILTER                          |                              |      1 |        |      0 |00:00:09.91 |    1301K|   

|*  3 |    HASH JOIN SEMI                 |                              |      1 |      1 |      0 |00:00:09.91 |    1301K|   

|   4 |     NESTED LOOPS                  |                              |      1 |      1 |     79 |00:00:10.64 |    1301K|   

|   5 |      NESTED LOOPS                 |                              |      1 |      1 |     79 |00:00:10.64 |    1301K|   

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

|   6 |       NESTED LOOPS                |                              |      1 |      5 |     85 |00:00:10.86 |    1300K|   

|*  7 |        TABLE ACCESS FULL          | T_PRODUCT_FEE                |      1 |     22 |  19202 |00:00:09.35 |    1262K|   

|*  8 |        TABLE ACCESS BY INDEX ROWID| T_CONTRACT_MASTER            |  19202 |      1 |     85 |00:00:00.42 | 

|*  9 |         INDEX UNIQUE SCAN         | PK_T_CONTRACT_MASTER         |  19202 |      1 |  19202 |00:00:00.18 |   19204 |

|* 10 |       TABLE ACCESS BY INDEX ROWID | T_POLICY_FEE                 |     85 |      1 |     79 |00:00:00.01 |     25

|* 11 |        INDEX UNIQUE SCAN          | PK_T_POLICY_FEE              |     85 |      1 |     85 |00:00:00.01 |     172 |   

|  12 |      TABLE ACCESS BY INDEX ROWID  | T_PRODUCT_LIFE               |     79 |      1 |     79 |00:00:00.01 |      81 |

|* 13 |       INDEX UNIQUE SCAN           | PK_T_PRODUCT_LIFE            |     79 |      1 |     79 |00:00:00.01 |       2 |   

|  14 |     VIEW                          | VW_NSO_1                     |      1 |     58 |     54 |00:00:00.01 |      23 |   

|* 15 |      FILTER                       |                              |      1 |        |     54 |00:00:00.01 |      23 |   

|* 16 |       CONNECT BY WITH FILTERING   |                              |      1 |        |     54 |00:00:00.01 |      23 |   

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

|  17 |        TABLE ACCESS BY INDEX ROWID| T_AGENCY                     |      1 |        |      1 |00:00:00.01 |       2 |   

|* 18 |         INDEX UNIQUE SCAN         | PK_T_AGENCY                  |      1 |      1 |      1 |00:00:00.01 |       1 |   

|* 19 |        HASH JOIN                  |                              |      3 |        |     53 |00:00:00.01 |      21 |   

|  20 |         CONNECT BY PUMP           |                              |      3 |        |     54 |00:00:00.01 |       0 |   

|  21 |         TABLE ACCESS FULL         | T_AGENCY                     |      3 |     58 |    177 |00:00:00.01 |      21 |   

|  22 |        TABLE ACCESS FULL          | T_AGENCY                     |      0 |     58 |      0 |00:00:00.01 |       0 |   

|* 23 |    INDEX RANGE SCAN               | UNI_LIFE_BASIC__PRODUCT_INFO |      0 |     39 |      0 |00:00:00.01 |       0 |   

----------------------------------------------------------------------------------------------------

 

Predicate Information (identified by operation id):

---------------------------------------------------

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

 

   2 - filter((TO_NUMBER("PL"."PERIOD_TYPE")2 OR  IS NOT NULL))

   3 - access("CM"."NB_AGENCY_ID"="$nso_col_1")

   7 - filter(("PF"."POLICY_YEAR"=2 AND INTERNAL_FUNCTION("PF"."FEE_TYPE") AND TO_NUMBER("PF"."CHARG

              "PKG_LIFE_CALC_BASIC_COMM"."F_IS_EFFECT"("PF"."LIST_ID")=0))

   8 - filter(("CM"."NB_AGENCY_ID" IS NOT NULL AND "CM"."VALIDATE_DATE"

              hh24:mi:ss') AND "CM"."PREM_STATUS"2 AND "CM"."VALIDATE_DATE">=TO_DATE('2009-03-01 00:00:00

              ADD_MONTHS(INTERNAL_FUNCTION("CM"."VALIDATE_DATE"),12)

              "CM"."PREM_STATUS"3 AND "CM"."PREM_STATUS"4))

   9 - access("PF"."POLICY_ID"="CM"."POLICY_ID")

  10 - filter((TO_NUMBER("PFE"."RECEIV_STATUS")=2 AND "PFE"."RECEIVED_TIME"

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

              hh24:mi:ss') AND "PFE"."RECEIVED_TIME">=ADD_MONTHS(INTERNAL_FUNCTION("CM"."VALIDATE_DATE"),12

  11 - access("PF"."FEE_ID"="PFE"."FEE_ID")

  13 - access("PF"."PRODUCT_ID"="PL"."PRODUCT_ID")

  15 - filter((TO_DATE('2009-04-01 00:00:00', 'yyyy-mm-dd hh24:mi:ss')>TO_DATE('2009-03-01 00:00:00'

              hh24:mi:ss') AND TO_DATE('2009-03-01 00:00:00', 'yyyy-mm-dd hh24:mi:ss')

              hh24:mi:ss')))

  16 - filter("AGENCY_ID"=121)

  18 - access("AGENCY_ID"=121)

  19 - access("PARENT_ID"=NULL)

  23 - access("CC"."PRODUCT_ID"=:B1)

       filter("CC"."COVERAGE_YEAR"1)

 

在进行第7步的时候花费了大量的时间,原因是,执行计划把PKG_LIFE_CALC_BASIC_COMM.F_IS_EFFECT这个筛选条件提到了语句的第一步执行,

这个时候返回的值很多,而每个返回的值都会去调用PKG_LIFE_CALC_BASIC_COMM.F_IS_EFFECT进行处理,所以语句1就会出现大量的重复操作,假设我这个时候返回了3W条那PKG_LIFE_CALC_BASIC_COMM.F_IS_EFFECT就会调用3W次,而语句1就会出现3W次,所以我们的问题转移到如何改变执行计划在最后一步来进行PKG_LIFE_CALC_BASIC_COMM.F_IS_EFFECT的处理。而这种等待既不会导致大量CPU计算,也不会导致大量I/O,因为进行的所有读均为逻辑读,也就是从内存进行读写,而ORACLE  SGA的值是恒定的2.5G左右,所以在最初看来,服务器性能视乎还是比较正常的。处于这些原因,我们开始考虑使用视图的方式来进行改写。

经过初步改写语句执行计划如下:

 

 

----------------------------------------------------------------------------------------------------

| Id  | Operation                         | Name                         | Starts | E-Rows | A-Rows |   A-Time   | Buffers |

----------------------------------------------------------------------------------------------------

|   1 |  SORT AGGREGATE                   |                              |      1 |      1 |      1 |00:00:09.32 |    1298K|

|*  2 |   FILTER                          |                              |      1 |        |      0 |00:00:09.32 |    1298K|

|*  3 |    HASH JOIN SEMI                 |                              |      1 |      1 |      0 |00:00:09.32 |    1298K|

|   4 |     NESTED LOOPS                  |                              |      1 |      1 |     79 |00:00:10.03 |    1298K|

|   5 |      NESTED LOOPS                 |                              |      1 |      1 |     79 |00:00:10.03 |    1298K|

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

|   6 |       NESTED LOOPS                |                              |      1 |      5 |     85 |00:00:10.24 |    1297K|

|*  7 |        TABLE ACCESS FULL          | T_PRODUCT_FEE                |      1 |     22 |  19155 |00:00:09.14 |    1259K|

|*  8 |        TABLE ACCESS BY INDEX ROWID| T_CONTRACT_MASTER            |  19155 |      1 |     85 |00:00:00.41 | 

|*  9 |         INDEX UNIQUE SCAN         | PK_T_CONTRACT_MASTER         |  19155 |      1 |  19155 |00:00:00.19 |   19157 |

|* 10 |       TABLE ACCESS BY INDEX ROWID | T_POLICY_FEE                 |     85 |      1 |     79 |00:00:00.01 |     25

|* 11 |        INDEX UNIQUE SCAN          | PK_T_POLICY_FEE              |     85 |      1 |     85 |00:00:00.01 |     172 |

|  12 |      TABLE ACCESS BY INDEX ROWID  | T_PRODUCT_LIFE               |     79 |      1 |     79 |00:00:00.01 |      81 |

|* 13 |       INDEX UNIQUE SCAN           | PK_T_PRODUCT_LIFE            |     79 |      1 |     79 |00:00:00.01 |       2 |

|  14 |     VIEW                          | VW_NSO_1                     |      1 |     58 |     54 |00:00:00.01 |      23 |

|* 15 |      FILTER                       |                              |      1 |        |     54 |00:00:00.01 |      23 |

|* 16 |       CONNECT BY WITH FILTERING   |                              |      1 |        |     54 |00:00:00.01 |      23 |

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

|  17 |        TABLE ACCESS BY INDEX ROWID| T_AGENCY                     |      1 |        |      1 |00:00:00.01 |       2 |

|* 18 |         INDEX UNIQUE SCAN         | PK_T_AGENCY                  |      1 |      1 |      1 |00:00:00.01 |       1 |

|* 19 |        HASH JOIN                  |                              |      3 |        |     53 |00:00:00.01 |      21 |

|  20 |         CONNECT BY PUMP           |                              |      3 |        |     54 |00:00:00.01 |       0 |

|  21 |         TABLE ACCESS FULL         | T_AGENCY                     |      3 |     58 |    177 |00:00:00.01 |      21 |

|  22 |        TABLE ACCESS FULL          | T_AGENCY                     |      0 |     58 |      0 |00:00:00.01 |       0 |

|* 23 |    INDEX RANGE SCAN               | UNI_LIFE_BASIC__PRODUCT_INFO |      0 |     39 |      0 |00:00:00.01 |       0 |

----------------------------------------------------------------------------------------------------

 

Predicate Information (identified by operation id):

---------------------------------------------------

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

 

   2 - filter((TO_NUMBER("PL"."PERIOD_TYPE")2 OR  IS NOT NULL))

   3 - access("CM"."NB_AGENCY_ID"="$nso_col_1")

   7 - filter(("PF"."POLICY_YEAR"=2 AND INTERNAL_FUNCTION("PF"."FEE_TYPE") AND TO_NUMBER("PF"."CHARG

              "PKG_LIFE_CALC_BASIC_COMM"."F_IS_EFFECT"("PF"."LIST_ID")=0))

   8 - filter(("CM"."NB_AGENCY_ID" IS NOT NULL AND "CM"."VALIDATE_DATE"

              hh24:mi:ss') AND "CM"."PREM_STATUS"2 AND "CM"."VALIDATE_DATE">=TO_DATE('2009-03-01 00:00:00

              hh24:mi:ss') AND ADD_MONTHS(INTERNAL_FUNCTION("CM"."VALIDATE_DATE"),12)

              hh24:mi:ss') AND "CM"."PREM_STATUS"3 AND "CM"."PREM_STATUS"4))

   9 - access("PF"."POLICY_ID"="CM"."POLICY_ID")

  10 - filter((TO_NUMBER("PFE"."RECEIV_STATUS")=2 AND "PFE"."RECEIVED_TIME"

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

              'yyyy-mm-dd hh24:mi:ss') AND "PFE"."RECEIVED_TIME">=ADD_MONTHS(INTERNAL_FUNCTION("CM"."VALIDA

  11 - access("PF"."FEE_ID"="PFE"."FEE_ID")

  13 - access("PF"."PRODUCT_ID"="PL"."PRODUCT_ID")

  15 - filter((TO_DATE('2009-04-01 00:00:00', 'yyyy-mm-dd hh24:mi:ss')>TO_DATE('2009-03-01 00:00:00'

              hh24:mi:ss') AND TO_DATE('2009-03-01 00:00:00', 'yyyy-mm-dd hh24:mi:ss')

              hh24:mi:ss')))

  16 - filter("AGENCY_ID"=121)

  18 - access("AGENCY_ID"=121)

  19 - access("PARENT_ID"=NULL)

  23 - access("CC"."PRODUCT_ID"=:B1)

       filter("CC"."COVERAGE_YEAR"1)

 

仔细分析后发现问题出在第7布,执行计划,并没有按照我们意图在进行所有的筛选条件过后再进行PKG_LIFE_CALC_BASIC_COMM.F_IS_EFFECT的处理,执行计划基本和以前一样。

   但是为了临时解决问题,我们通过建立临时中间表的方法来解决这个问题,对子查询建立临时表,强制改变其访问路径。但是我们同时也考虑到,建立临时表毕竟需要对临时表进行频繁的INSERTDELETE,所以最好的方式还是改写SQL本身,改变其访问路径。

   所以经过我仔细分析后,希望通过改写语句如下:

select    decode(sum(period_prem), null, 0, sum(period_prem)) from

     (select  decode(pkg_life_calc_basic_comm.f_is_effect(pf.list_id),'0',PF.list_id) list_id, decode(pkg_life_calc_basic_comm.f_is_effect(PF.list_id),'0',nvl(pf.period_prem, 0) )period_prem

                          from t_product_fee     pf,

                                t_contract_master cm,

                                t_product_life    pl,

                                t_policy_fee      pfe

                          where pf.policy_id = cm.policy_id

                            and cm.validate_date

                               ADD_MONTHS(to_date('201005', 'yyyymm'), -13)

                           and cm.validate_date >=

                               ADD_MONTHS(to_date('201005', 'yyyymm'), -14)

                           and pf.fee_id = pfe.fee_id

                           and pfe.received_time >=

                               ADD_MONTHS(cm.validate_date, 12)

                           and pfe.received_time

                               ADD_MONTHS(to_date('201005', 'yyyymm'), 1)

                           and pfe.receiv_status = 2

                           and pf.charge_type = 1

                           and pf.product_id = pl.product_id

                           and (pl.period_type 2 or exists

                                (select 1

                                  from t_life_basic cc

                                  where cc.coverage_year 1

                                    and cc.product_id = pl.product_id))

                           and pf.fee_type in (43, 47, 48)

                          and pf.policy_year = 2

                          and cm.prem_status not in (2, 3, 4)

                           and cm.nb_agency_id in

                              (select    agency_id

                                from t_agency

                                 start with agency_id = 121

                                Connect By Prior agency_id = parent_id));

如红色部分使用DECODE语句来改写WHERE语句,再次查看执行计划如下:

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

 

Plan hash value: 4085472284

 

----------------------------------------------------------------------------------------------------

| Id  | Operation                        | Name                         | Starts | E-Rows | A-Rows |   A-Time   | Buffers |

----------------------------------------------------------------------------------------------------

|   1 |  SORT AGGREGATE                  |                              |      1 |      1 |      1 |00:00:00.29 |   20999 |

|*  2 |   FILTER                         |                              |      1 |        |      0 |00:00:00.29 |   20999 |

|   3 |    NESTED LOOPS                  |                              |      1 |      5 |   2019 |00:00:00.25 |   20941 |

|   4 |     NESTED LOOPS                 |                              |      1 |      5 |   2019 |00:00:00.21 |   18920 |

|   5 |      NESTED LOOPS                |                              |      1 |    101 |   2137 |00:00:00.17 |   12507 |

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

|*  6 |       TABLE ACCESS FULL          | T_CONTRACT_MASTER            |      1 |    100 |   2749 |00:00:00.03 |   

|*  7 |       TABLE ACCESS BY INDEX ROWID| T_PRODUCT_FEE                |   2749 |      1 |   2137 |00:00:00.10 |    80

|*  8 |        INDEX RANGE SCAN          | IDX_PRODUCT_FEE__POLICY_ID   |   2749 |      2 |   7161 |00:00:00.03 |

|*  9 |      TABLE ACCESS BY INDEX ROWID | T_POLICY_FEE                 |   2137 |      1 |   2019 |00:00:00.05 |    64

|* 10 |       INDEX UNIQUE SCAN          | PK_T_POLICY_FEE              |   2137 |      1 |   2137 |00:00:00.02 |    4276 |

|  11 |     TABLE ACCESS BY INDEX ROWID  | T_PRODUCT_LIFE               |   2019 |      1 |   2019 |00:00:00.03 |    2

|* 12 |      INDEX UNIQUE SCAN           | PK_T_PRODUCT_LIFE            |   2019 |      1 |   2019 |00:00:00.01 |       2 |

|* 13 |    INDEX RANGE SCAN              | UNI_LIFE_BASIC__PRODUCT_INFO |      6 |     39 |      0 |00:00:00.01 | 

|* 14 |     FILTER                       |                              |      2 |        |      0 |00:00:00.01 |      46 |

|* 15 |      CONNECT BY WITH FILTERING   |                              |      2 |        |    108 |00:00:00.01 |      46 |

|  16 |       TABLE ACCESS BY INDEX ROWID| T_AGENCY                     |      2 |        |      2 |00:00:00.01 |       4 |

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

|* 17 |        INDEX UNIQUE SCAN         | PK_T_AGENCY                  |      2 |      1 |      2 |00:00:00.01 |       2 |

|* 18 |       HASH JOIN                  |                              |      6 |        |    106 |00:00:00.01 |      42 |

|  19 |        CONNECT BY PUMP           |                              |      6 |        |    108 |00:00:00.01 |       0 |

|  20 |        TABLE ACCESS FULL         | T_AGENCY                     |      6 |     58 |    354 |00:00:00.01 |      42 |

|  21 |       TABLE ACCESS FULL          | T_AGENCY                     |      0 |     58 |      0 |00:00:00.01 |       0 |

----------------------------------------------------------------------------------------------------

 

Predicate Information (identified by operation id):

---------------------------------------------------

 

   2 - filter(((TO_NUMBER("PL"."PERIOD_TYPE")2 OR  IS NOT NULL) AND  IS NOT NULL))

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

   6 - filter(("CM"."VALIDATE_DATE"

              "CM"."PREM_STATUS"2 AND "CM"."VALIDATE_DATE">=TO_DATE('2009-03-01 00:00:00', 'yyyy-mm-dd hh

              ADD_MONTHS(INTERNAL_FUNCTION("CM"."VALIDATE_DATE"),12)

              "CM"."PREM_STATUS"3 AND "CM"."PREM_STATUS"4))

   7 - filter(("PF"."POLICY_YEAR"=2 AND INTERNAL_FUNCTION("PF"."FEE_TYPE") AND TO_NUMBER("PF"."CHARG

   8 - access("PF"."POLICY_ID"="CM"."POLICY_ID")

   9 - filter((TO_NUMBER("PFE"."RECEIV_STATUS")=2 AND "PFE"."RECEIVED_TIME"

              'yyyy-mm-dd hh24:mi:ss') AND "PFE"."RECEIVED_TIME">=ADD_MONTHS(INTERNAL_FUNCTION("CM"."VALIDA

  10 - access("PF"."FEE_ID"="PFE"."FEE_ID")

  12 - access("PF"."PRODUCT_ID"="PL"."PRODUCT_ID")

  13 - access("CC"."PRODUCT_ID"=:B1)

 

PLAN_TABLE_OUTPUT

----------------------------------------------------------------------------------------------------

       filter("CC"."COVERAGE_YEAR"1)

  14 - filter("AGENCY_ID"=:B1)

  15 - filter("AGENCY_ID"=121)

  17 - access("AGENCY_ID"=121)

  18 - access("PARENT_ID"=NULL)

此时已经第7步没有进行方法处理,这个处理会再所有的谓词过滤后执行。这个语句的执行时间为0.3秒左右,比起最开始的15秒得到了大量的提升

 

二、解决方法

所以这个问题的解决方法如下:

1、改写SQL语句如下:

select /*+  gather_plan_statistics */

   decode(sum(period_prem), null, 0, sum(period_prem)) from

     (select  decode(pkg_life_calc_basic_comm.f_is_effect(pf.list_id),'0',PF.list_id) list_id, decode(pkg_life_calc_basic_comm.f_is_effect(PF.list_id),'0',nvl(pf.period_prem, 0) )period_prem

                          from t_product_fee     pf,

                                t_contract_master cm,

                                t_product_life    pl,

                                t_policy_fee      pfe

                          where pf.policy_id = cm.policy_id

                            and cm.validate_date

                               ADD_MONTHS(to_date('201005', 'yyyymm'), -13)

                           and cm.validate_date >=

                               ADD_MONTHS(to_date('201005', 'yyyymm'), -14)

                           and pf.fee_id = pfe.fee_id

                           and pfe.received_time >=

                               ADD_MONTHS(cm.validate_date, 12)

                           and pfe.received_time

                               ADD_MONTHS(to_date('201005', 'yyyymm'), 1)

                           and pfe.receiv_status = 2

                           and pf.charge_type = 1

                           and pf.product_id = pl.product_id

                           and (pl.period_type 2 or exists

                                (select 1

                                  from t_life_basic cc

                                  where cc.coverage_year 1

                                    and cc.product_id = pl.product_id))

                           and pf.fee_type in (43, 47, 48)

                          and pf.policy_year = 2

                          and cm.prem_status not in (2, 3, 4)

                           and cm.nb_agency_id in

                              (select    agency_id

                                from t_agency

                                 start with agency_id = 121

                                Connect By Prior agency_id = parent_id));

2、考虑到我们的系统session_cached_cursors过小,同时引起了一些等待,所以尽可能更改参数session_cached_cursors加大每个会话能够缓存CURSOR的数量。

相关文章
|
7月前
对象的优化
对象的优化
|
2月前
|
缓存 Python
深度解密为什么实例在调用方法时会将自身传给 self 参数(一)
深度解密为什么实例在调用方法时会将自身传给 self 参数
46 0
|
2月前
|
设计模式 Python
深度解密为什么实例在调用方法时会将自身传给 self 参数(二)
深度解密为什么实例在调用方法时会将自身传给 self 参数(二)
39 1
|
7月前
|
网络协议 C++
一文彻底解决C++中的重载、重写和隐藏操作
一文彻底解决C++中的重载、重写和隐藏操作
90 0
通过普通方式调用和反射方式调用以及关闭检测后,通过反射方式调用的性能对比
通过普通方式调用和反射方式调用以及关闭检测后,通过反射方式调用的性能对比
|
Java 编译器
深入探索: 对象构造的隐藏功能与技巧
深入探索: 对象构造的隐藏功能与技巧
101 0
|
C# 编译器
C#隐藏方法
不能删除基类的任何成员,但可以用与基类成员相同的成员来屏蔽基类成员 屏蔽数据成员:派生类中声明名称和类型相同的成员 屏蔽函数成员:在派生类中声明新的带有函数签名的成员 让编译器知道:添加new关键字,否则会警告pet petpu.
705 0
动态隐藏行
动态隐藏行
853 0
|
数据安全/隐私保护