SQL字符串截取

今天接到一个需求，统计一下邮箱系统一至九月分的用户开销户记录，开户记录很容易就查到了，但由于用户销户后数据库表中找不到对应的用户信息，查了一系列的表，发现SBMailSubUserLog其中一列中包含有用户的部分信息，如果想要使用的话必须截取，然后我们步入正题，截取数据串。

原始的字符串‘邮箱成功开通

Passwd=Nm,12345,Mailsize=30,Mailname=gtxx_lijie,Displayname=李洁,saa’

现在我们把Mailname和Displayname提取出来，以Displayname为例，大体思路是这样：把‘Displayname=’和前面的数据截取掉，留下后面的数据，

SQL：

Select SUBSTRING(subdesc,CHARINDEX('DISPLAYNAME',subdesc)+12,LEN(subdesc)-CHARINDEX('DISPLAYNAME',subdesc)) from SBMailSubUserLog where subtype like '%员工业务开通%' and mobile in (select distinct mobile from SBMailSubUserLog where subtype like '%员工业务关闭%' and changedate>='2009-01-01 00:32:24.327')

显示出来的就是这样’ 李洁,saa’现在再整理‘，’号之后的数据select left(subdesc，patindex('%,%', subdesc）） from SBMailSubUserLog，然后把它们整合起来，

Select left(SUBSTRING(subdesc,CHARINDEX('DISPLAYNAME',subdesc)+12,LEN(subdesc)-CHARINDEX('DISPLAYNAME',subdesc)),patindex('%,%',SUBSTRING(subdesc,CHARINDEX('DISPLAYNAME',subdesc)+12,LEN(subdesc)-CHARINDEX('DISPLAYNAME',subdesc))))

 from SBMailSubUserLog

where subtype like '%员工业务开通%' and

 mobile in

(select distinct mobile from SBMailSubUserLog where subtype like '%员工业务关闭%' and changedate>='2009-01-01 00:32:24.327')。

完整的提取数据的SQL：

select distinct a.mobile,b.displayname,b.mailname,

datepart(year,a.changedate)*10000+datepart(month,a.changedate)*100+datepart(day,a.changedate) as changedate2,

a.source,a.subtype,a.subdesc

from SBMailSubUserLog a join

(select mobile,

left(SUBSTRING(subdesc,CHARINDEX('DISPLAYNAME',subdesc)+12,LEN(subdesc)-CHARINDEX('DISPLAYNAME',subdesc)),patindex

('%,%',SUBSTRING(subdesc,CHARINDEX('DISPLAYNAME',subdesc)+12,LEN(subdesc)-CHARINDEX('DISPLAYNAME',subdesc)))) as displayname,

left(SUBSTRING(subdesc,CHARINDEX('mailname',subdesc)+12,LEN(subdesc)-CHARINDEX('mailname',subdesc)),patindex

('%,%',SUBSTRING(subdesc,CHARINDEX('mailname',subdesc)+12,LEN(subdesc)-CHARINDEX('mailname',subdesc)))) as mailname,

source,subtype,datepart(year,changedate)*10000+datepart(month,changedate)*100+datepart(day,changedate) as changedate1

 from SBMailSubUserLog

where subtype like '%员工业务开通%' and

 mobile in

(select distinct mobile from SBMailSubUserLog where subtype like '%员工业务关闭%' and changedate>='2009-01-01 00:32:24.327')

)  b on a.mobile=b.mobile

where a.subtype like '%员工业务关闭%'

--and a.changedate>='2009-01-01 00:32:24.327'

and datepart(year,a.changedate)*100+datepart(month,a.changedate)='200904'

order by changedate2