t-sql 0 0~

[研小艾]

select ( repair.rgroup ),count(1) from repair,BPMInstTasks where repair.taskid=BPMInstTasks.taskid and BPMInstTasks.state='running' group by repair.rgroup

BPMInstTasks 为流程表

repair为某流程表

1.现在想加上满足BPMInstTasks.state='running' 的不分组的count
group count
group1 3
group2 2
group3 4
all 9

[cd731107]

select ( repair.rgroup ),count(1) from repair,BPMInstTasks 
where repair.taskid=BPMInstTasks.taskid and BPMInstTasks.state='running' 
group by repair.rgroup
union all
select 'all',count(1) from repair,BPMInstTasks 
where repair.taskid=BPMInstTasks.taskid and BPMInstTasks.state='running'

[研小艾]

to 豆子:
你觉得是我想要的??

[cd731107]

是这么理解的：用下面这句
select 'all',count(1) from repair,BPMInstTasks
where repair.taskid=BPMInstTasks.taskid and BPMInstTasks.state='running'
表示最后的那条记录 all 9

[研小艾]

0 0~ 一条select

[cd731107]

select isnull(( repair.rgroup ),'all'),count(1) from repair,BPMInstTasks 
where repair.taskid=BPMInstTasks.taskid and BPMInstTasks.state='running' 
group by repair.rgroup
WITH ROLLUP

[-晴天]

既然是不分组,怎么会有你要的这样的结果呢?
/*
现在想加上满足BPMInstTasks.state='running' 的不分组的count
group count
group1 3
group2 2
group3 4
all 9
*/
group 是什么东西?

[研小艾]

to 晴天:
如果我们的报表程序支持2個select 那么[豆子]
的可以

就是分组查询之后加上一行总的

[cd731107]

--或使用cube
select isnull(( repair.rgroup ),'all'),count(1) from repair,BPMInstTasks 
where repair.taskid=BPMInstTasks.taskid and BPMInstTasks.state='running' 
group by repair.rgroup
WITH cube

[研小艾]

有木有t-sql 的support 0 0~
求t-sql的technology support 一枚就好0 0~

[研小艾]

结贴不解释0 0~ 请继续关注 还有各种2B的需求 您就乖巧的拿分吧0 0~