They say at Christmas it is better to give than to receive.

So I will give you back your sample data in a manner that is consumable and thereby the helpers on this forum can give you better help.

WITH SampleData (Userid, [DateTime], [Check-In-Type]) AS

(

SELECT '001',CAST('2013-01-01 12:30' AS DATETIME),'I'

UNION ALL SELECT '001','2013-01-01 12:31','I'

UNION ALL SELECT '001','2013-01-01 12:35','I'

UNION ALL SELECT '002','2013-01-01 12:30','I'

UNION ALL SELECT '002','2013-01-01 12:31','I'

UNION ALL SELECT '002','2013-01-01 12:35','I'

UNION ALL SELECT '002','2013-01-01 12:31','O'

UNION ALL SELECT '002','2013-01-01 12:35','O'

)

SELECT Userid

,DateOfCheckInOut=DATEADD(day, DATEDIFF(day, 0, [DateTime]), 0)

,MinCheckIn=MIN(CASE [Check-In-Type] WHEN 'I' THEN [DateTime] END)

,MaxCheckOut=MAX(CASE [Check-In-Type] WHEN 'O' THEN [DateTime] END)

FROM SampleData

GROUP BY DATEADD(day, DATEDIFF(day, 0, [DateTime]), 0), Userid;

I've also proposed a solution that may be close to what you're looking for. As Chris said though, you need to be more clear on your business rules before we can be sure of what you're trying to do.

In particular, I was quite unsure of the data types of your columns, so I converted the Date/Time to a single column of DATETIME data type. If yours are truly separate, you can combine them and use what I gave you. Better if your table already has them combined. Keeping them separate will just cause you no end to headaches later on.

Merry Christmas!