care to share your table scripts / indexes / sample data scripts that demonstrates your issue?

you seem to be making several postings to very old threads......????

this looks like MySQL code ??

Hi Jeff...appreciate you comment.

however been playing around...always dangerous...

ok..try this...far from elegant, and performance may be poor dependent upon number of rows and indexes

hopefully some else will chime in with a better idea.....

SELECT DISTINCT ClientId

FROM...

so can I assume ClientId 3 qualifies as well?

SELECT ClientId, ActivityId, CreateDate,...

why doesn't ClientId "1" qualify?

SELECT ClientId, ActivityId, CreateDate, DATEDIFF(d, CreateDate, GETDATE()) AS datediff

FROM TestPurgeClient

WHERE (ClientId = 1)

can we start again with a bigger data set that demonstrates all your possibilities and the expected outcome.

sorry...but i'm obviously not thinking clearly 😛

New Person (5/18/2013)

---

Thank you very much for your kind response. For this case it is working fine. But in this SQL activity ID is not taken into consideration at all...

normally formatting is better left to the front end app to perform.

sometimes however this isn't always poss or desired.....

so in this case, how does your client want the following displayed?

-10,000,000

1

10

100

1,000

10,000

100,000

1,000,000

as a start...try this

SELECT ClientId

FROM TestPurgeClient

GROUP BY ClientId

HAVING (SUM(CASE WHEN DATEDIFF(d, CreateDate, GETDATE()) > 90 THEN 1 ELSE 0 END) =...

AndrewSQLDBA (5/14/2013)

---

I have not been able to every find that SSMS will do that. I was hoping that someone had found a way.

I have the SSMS Tools. So how would...

http://www.ssmstoolspack.com/Licensing

take a look above...still free for 2008.

don't think that SSMS will do what you want...but be very pleased to be proved wrong .

as a guide

SELECT COUNT(employee) AS All_cnt,

sum(case when hiredate = '1995' then 1 else 0 end) as HY_1995 ,

sum(case when hiredate = '1996' then 1 else 0 end)...

Matthew Cushing (5/13/2013)

---

J Livingston SQL (5/13/2013)

personally I would model this in excel and pivot tables first.....easy enough to create sample data.

start planning on what you really want to analyse /...