客户提出需求，针对某一列分组加上小计，合计汇总。网上找了一些有关SQL加合计的语句。都不是很理想。决定自己动手写。 思路有三个： 1.很多用GROUPPING和ROLLUP来实现。 优点：实现
客户提出需求，针对某一列分组加上小计，合计汇总。网上找了一些有关SQL加合计的语句。都不是很理想。决定自己动手写。思路有三个：1.很多用GROUPPING和ROLLUP来实现。 优点：实现代码简洁，要求对GROUPPING和ROLLUP很深的理解。 缺点：低版本的Sql Server不支持。2.游标实现。 优点：思路逻辑简洁。 缺点：复杂和低效。3.利用临时表。 优点：思路逻辑简洁，执行效率高。SQL实现简单。 缺点：数据量大时耗用内存.
综合三种情况，决定“利用临时表”实现。实现效果原始表TB
加上小计，合计后效果
SQL语句复制代码 代码如下:select * into #TB from TB select * into #TB1 from #TB where 1<>1 select distinct zcxt into #TBype from #TB order by zcxtselect identity(int,1,1) fid,zcxt into #TBype1 from #TBypeDECLARE @i intDECLARE @k int select @i=COUNT(*) from #TBype set @k=0 DECLARE @strfname varchar(50) WHILE @k < @i BEGIN Set @k =@k +1 select @strfname=zcxt from #TBype1 where fid =@k set IDENTITY_INSERT #TB1 ON insert into #TB1(fid,qldid,fa_cardid,ztbz,fa_name,model,i_number,gzrq,zcyz,ljzj,jz,sybm,zcxt,fa_ljjzzb) select fid,qldid,fa_cardid,ztbz,fa_name,model,i_number,gzrq,zcyz,ljzj,jz,sybm,zcxt,fa_ljjzzb from ( select * from #TB where zcxt=@strfname union all select 0 fid,'' qldid,'' fa_cardid,'' ztbz,'小计' fa_name,'' model,sum(i_number) as i_number,'' gzrq,sum(CAST(zcyz as money)) as zcyz,sum(CAST(ljzj as money)) as ljzj,sum(CAST(jz as money)) as jz,'' sybm,'' zcxt,Sum(fa_ljjzzb) as fa_ljjzzb from #TB where zcxt=@strfname group by ztbz ) as B set IDENTITY_INSERT #TB1 off ENDselect qldid,fa_cardid,zcxt,fa_name,model,i_number,gzrq,zcyz,ljzj,jz,sybm,ztbz,fa_ljjzzb from #TB1union allselect '' qldid,'' fa_cardid,'' ztbz,'合计' fa_name,'' model,sum(i_number) as i_number,'' gzrq,sum(CAST(zcyz as money)) as zcyz,sum(CAST(ljzj as money)) as ljzj,sum(CAST(jz as money)) as jz,'' sybm,'' zcxt,Sum(fa_ljjzzb) as fa_ljjzzbfrom #TBdrop table #TB1 drop table #TBype1drop table #TBypedrop table #TB 扩展改进可以改写成一个通用的添加合计小计的存储过程。
