一条长sql的排错过程

过程

有这样一条长sql，由于环境原因，对select.....in...... 语法限制使用，因此以left  join语法代替，原来只需要统计一天的数据汇总结果，因为特殊需要，需要一次性统计各天的结果，于是首先将子查询中时间过滤的部分提取到外层。

修改前sql1：

select count(*) from (SELECT DISTINCT t.con_no AS con_no

   , CASE 

WHEN tgua.atime IS NOT NULL THEN tgua.atime

WHEN tcc.CTIME IS NOT NULL THEN tcc.CTIME

WHEN bb.CTIME IS NOT NULL THEN bb.CTIME

 END AS CTIME, t.PRO_METHOD AS pro_method

FROM my_customer t

LEFT OUTER JOIN first_ auth tgua

ON t.con_no = tgua.con_no

AND tgua.stat = 1

AND tgua.type = 'C' and 

(tgua.atime>= '2017-06-22' and tgua.atime< '2017-06-23')

LEFT OUTER JOIN my_ contact tcc

ON t.con_no = tcc.con_no

AND tcc.NAME IS NOT NULL and 

(tcc.ctime>= '2017-06-22' and tcc.ctime< '2017-06-23')

LEFT OUTER JOIN (

SELECT tcp.con_no AS con_no, tcai.CTIME AS CTIME

FROM my_ person tcp

INNER JOIN my_auth_info tcai

ON tcp.CNO = tcai.CNO and 

(tcai.ctime>= '2017-06-22' and tcai.ctime< '2017-06-23')

) bb

ON t.con_no = bb.con_no 

WHERE t.PRO_METHOD = 'uc' and (tgua.con_no IS NOT NULL

OR tcc.con_no IS NOT NULL

OR bb.con_no IS NOT NULL)) pp order by pp.con_no

sql2：

select count(*) from 

(SELECT DISTINCT t.con_no AS con_no

, CASE 

                               WHEN tgua.atime IS NOT NULL THEN tgua.atime

            WHEN tcc.CTIME IS NOT NULL THEN tcc.CTIME

            WHEN bb.CTIME IS NOT NULL THEN bb.CTIME

                        END AS CTIME, t.PRO_METHOD AS pro_method

FROM my_customer t

LEFT OUTER JOIN first_ auth tgua

ON t.con_no = tgua.con_no

AND tgua.status = 1

AND tgua.type = 'C'

LEFT OUTER JOIN my_ contact tcc

ON t.con_no = tcc.con_no

AND tcc.NAME IS NOT NULL

LEFT OUTER JOIN (

SELECT tcp.con_no AS con_no, tcai.CTIME AS CTIME

FROM my_ person tcp

INNER JOIN my_auth_info tcai

ON tcp.CNO = tcai.CNO

) bb

ON t.con_no = bb.con_no

WHERE t.pro_method = 'uc' and (tgua.con_no IS NOT NULL

OR tcc.con_no IS NOT NULL

OR bb.con_no IS NOT NULL)) p where CTIME>= '2017-06-22' and CTIME< '2017-06-23'

order by con_no 

试着执行了两条sql，发现sql2查询结果数量总比sql1要少，于是用*代替count(*),打印出具体查询结果，

sql3：

select * from (SELECT DISTINCT t.con_no AS con_no

, CASE 

WHEN tgua.atime IS NOT NULL THEN tgua.atime

WHEN tcc.CTIME IS NOT NULL THEN tcc.CTIME

WHEN bb.CTIME IS NOT NULL THEN bb.CTIME

END AS CTIME, t.PRO_METHOD AS pro_method

    FROM my_customer t

LEFT OUTER JOIN first_ auth tgua

ON t.con_no = tgua.con_no

AND tgua.stat = 1

AND tgua.type = 'C' and 

(tgua.atime>= '2017-06-22' and tgua.atime< '2017-06-23')

LEFT OUTER JOIN my_ contact tcc

ON t.con_no = tcc.con_no

AND tcc.NAME IS NOT NULL and 

(tcc.ctime>= '2017-06-22' and tcc.ctime< '2017-06-23')

LEFT OUTER JOIN (

SELECT tcp.con_no AS con_no, tcai.CTIME AS CTIME

FROM my_ person tcp

INNER JOIN my_auth_info tcai

ON tcp.CNO = tcai.CNO and 

(tcai.ctime>= '2017-06-22' and tcai.ctime< '2017-06-23')

) bb

ON t.con_no = bb.con_no 

WHERE t.PRO_METHOD = 'uc' and (tgua.con_no IS NOT NULL

OR tcc.con_no IS NOT NULL

OR bb.con_no IS NOT NULL)) pp order by pp.con_no

sql4：

select * from 

(SELECT DISTINCT t.con_no AS con_no

, CASE 

                             WHEN tgua.atime IS NOT NULL THEN tgua.atime

WHEN tcc.CTIME IS NOT NULL THEN tcc.CTIME

WHEN bb.CTIME IS NOT NULL THEN bb.CTIME

    END AS CTIME, t.PRO_METHOD AS pro_method

FROM my_customer t

LEFT OUTER JOIN first_ auth tgua

ON t.con_no = tgua.con_no

AND tgua.status = 1

AND tgua.type = 'C'

LEFT OUTER JOIN my_ contact tcc

ON t.con_no = tcc.con_no

AND tcc.NAME IS NOT NULL

LEFT OUTER JOIN (

SELECT tcp.con_no AS con_no, tcai.CTIME AS CTIME

FROM my_ person tcp

INNER JOIN my_auth_info tcai

ON tcp.CNO = tcai.CNO

) bb

ON t.con_no = bb.con_no

WHERE t.pro_method = 'uc' and (tgua.con_no IS NOT NULL

OR tcc.con_no IS NOT NULL

OR bb.con_no IS NOT NULL)) p where CTIME>= '2017-06-22' and CTIME< '2017-06-23'

order by con_no

找出sql3比sql4多出的执行结果记录，如编号为4534的一条记录, 这条记录在sql4中很有可能被过滤掉了，因此修改sql4外层的where查询条件，如下：

select * from 

(SELECT DISTINCT t.con_no AS con_no

, CASE 

 WHEN tgua.atime IS NOT NULL THEN tgua.atime

   WHEN tcc.CTIME IS NOT NULL THEN tcc.CTIME

   WHEN bb.CTIME IS NOT NULL THEN bb.CTIME

 END AS CTIME, t.PRO_METHOD AS pro_method

 FROM my_customer t

LEFT OUTER JOIN first_ auth tgua

ON t.con_no = tgua.con_no

AND tgua.status = 1

AND tgua.type = 'C'

LEFT OUTER JOIN my_ contact tcc

ON t.con_no = tcc.con_no

AND tcc.NAME IS NOT NULL

LEFT OUTER JOIN (

SELECT tcp.con_no AS con_no, tcai.CTIME AS CTIME

FROM my_ person tcp

INNER JOIN my_auth_info tcai

ON tcp.CNO = tcai.CNO

) bb

ON t.con_no = bb.con_no

WHERE t.pro_method = 'uc' and (tgua.con_no IS NOT NULL

OR tcc.con_no IS NOT NULL

OR bb.con_no IS NOT NULL)) p where con_no='4534'

查询结果发现CTIME对应的值竟是“2017-06-23 00:18:07“。而在sql3执行结果中，CTIME值为2017-06-22 00:16:08。问题出在这一sql片段中：

        CASE 

            WHEN tgua.atime IS NOT NULL THEN tgua.atime

  WHEN tcc.CTIME IS NOT NULL THEN tcc.CTIME

             WHEN bb.CTIME IS NOT NULL THEN bb.CTIME

          END AS CTIME, t.PRO_METHOD AS pro_method

在sql3的执行结果中，由于时间筛选在内层的left join后面，时间不是2017-06-22的记录字段值都用null代替，因此在结果集中，无论是tgua.atime，tcc.CTIME 还是bb.CTIME，时间不是null就是2017-06-22 ，而在sql4中，tgua.atime、tcc.CTIME 、bb.CTIME的时间是不一致的，因此假如case when选择一个not null的时间值，且该时间刚好不是2017-06-22，那么这条记录会在外层的where筛选中被过滤掉。而我们希望的却是tgua.atime、tcc.CTIME 、bb.CTIME只要有一个时间是在2017-06-22，那这条记录就应该出现在结果集中。使用union语法，修改如下：

select distinct * from 

((SELECT DISTINCT t.con_no AS con_no 

, DATE_FORMAT(tgua.atime,"%Y-%m-%d") AS ctime, t.PRO_METHOD AS pro_method

FROM my_customer t

inner JOIN first_ auth tgua

ON t.con_no = tgua.con_no

AND tgua.status = 1

AND tgua.type = 'C')

union

(SELECT DISTINCT t.con_no AS con_no 

, DATE_FORMAT(tcc.ctime,"%Y-%m-%d") AS ctime, t.PRO_METHOD AS pro_method

FROM my_customer t

inner JOIN

my_ contact tcc

ON t.con_no = tcc.con_no

AND tcc.NAME IS NOT NULL)

union

(SELECT DISTINCT t.con_no AS con_no 

,DATE_FORMAT(bb.ctime,"%Y-%m-%d") AS ctime, t.PRO_METHOD AS pro_method

FROM my_customer t

inner JOIN (

SELECT tcp.con_no AS con_no, tcai.ctime AS ctime

FROM my_ person tcp

INNER JOIN my_auth_info tcai

ON tcp.CNO = tcai.CNO

) bb

ON t.con_no = bb.con_no)) p where pro_method = 'uc' and ctime>= '2017-06-22' and ctime< '2017-06-23'

order by con_no 

查出的记录数恰好与sql1中的一致，在此基础上进行group by分组统计处理。

总结

     多表关联查询排错遇到问题时，要通过select *打印出具体的明细，进行对比，分析其特征，再反观sql思考其错误的可能原因，再修改，再排查，这是一个不断从现象到本质，从思考到实践的循环上升的过程，这也是修改一切bug的要旨所在。另外，越复杂的sql越容易出错，而且不利于修改维护，因此，在应用层编写代码时，sql越简单越好。

参考文章：mysql left( right ) join使用on 与where 筛选的差异

java达人

ID：java_daren

（长按或扫码识别）

                   

友情推荐

虾说十话