Again your sample data doesn't have any duplicates. The output you displayed is exactly what's in the sample data.

For my solution, I assumed Team ID was not required or that it was just a typo of the data. If my assumptions were both wrong, then yes, we...

I believe you could just group the information:

select sno, name, case, max(datetime) as MaxDate

from tablename

group by sno, name, case

oops...i just noticed the last line of your first post.

sorry about that.

why would you do this? I would think you would just change the compatibility of the database that needs SQL 2005 to SQL 2005.

are you running them on the same database server or is one test and the other production?

same answer as your other post. Look at Rank() or row_number().

Look at using the Rank() function and select a RanK() <= 3.

fail

run profiler using the locks counters. There is a default TSQL_Locks template...try that. You should get your sql in question.

To answer your question: "is this design OK?"

Yes and No. There are different approaches to storing historial data and different needs on accessing historical data. Not...

I think you are asking for a SELECT that changes the values of the data during the SELECT process, but does not update the actual table....correct? If that is...

ok then...do this... (use the query i sent you as a subquery...then collapse the columns) -- typing as i think

SELECT DataType,

MAX(CASE

WHEN DataType in ('CURRENCY','NUMBER') THEN CAST(MaxNumer as varchar)

WHEN...

use the dbcc sqlperf(logspace) command...

if you want to monitor it, create a table and insert the results from the command into the table.

First off...yes this is the WRONG way to do things.

You won't be able to do it in a single field, but you can do it with multiple fields...