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：
[align=left]select distinct a.mobile,b.displayname,b.mailname,[/align]
[align=left]datepart(year,a.changedate)*10000+datepart(month,a.changedate)*100+datepart(day,a.changedate) as changedate2,[/align]
[align=left]a.source,a.subtype,a.subdesc[/align]
[align=left]from SBMailSubUserLog a join [/align]
[align=left](select mobile,[/align]
[align=left]left(SUBSTRING(subdesc,CHARINDEX('DISPLAYNAME',subdesc)+12,LEN(subdesc)-CHARINDEX('DISPLAYNAME',subdesc)),patindex[/align]
[align=left]('%,%',SUBSTRING(subdesc,CHARINDEX('DISPLAYNAME',subdesc)+12,LEN(subdesc)-CHARINDEX('DISPLAYNAME',subdesc)))) as displayname,[/align]
[align=left]left(SUBSTRING(subdesc,CHARINDEX('mailname',subdesc)+12,LEN(subdesc)-CHARINDEX('mailname',subdesc)),patindex[/align]
[align=left]('%,%',SUBSTRING(subdesc,CHARINDEX('mailname',subdesc)+12,LEN(subdesc)-CHARINDEX('mailname',subdesc)))) as mailname,[/align]
[align=left]source,subtype,datepart(year,changedate)*10000+datepart(month,changedate)*100+datepart(day,changedate) as changedate1[/align]
[align=left] from SBMailSubUserLog[/align]
[align=left]where subtype like '%员工业务开通%' and[/align]
[align=left] mobile in[/align]
[align=left](select distinct mobile from SBMailSubUserLog where subtype like '%员工业务关闭%' and changedate>='2009-01-01 00:32:24.327')[/align]
[align=left]) b on a.mobile=b.mobile[/align]
[align=left]where a.subtype like '%员工业务关闭%' [/align]
[align=left]--and a.changedate>='2009-01-01 00:32:24.327' [/align]
[align=left]and datepart(year,a.changedate)*100+datepart(month,a.changedate)='200904'[/align]
order by changedate2本文出自 “逸志斋” 博客，转载请与作者联系！