And to add to what Rod said, if you have relatively few rows with week_calculated = 0, then make it a tinyint and index on that. At the very least, you need a clustered index ( you can only have one). If you most often search on the date field, then that may be a good candidate. Also - keep in mind that just because you are only returning 1000 rows, SQL Server in your case, will still need to do a full table scan to pick off which would be the top 1000 rows.





>The table has 10 million rows. The query is
>select top 1000 * from sales where date > '1/27/2003' and week_calculated = 0
>
>date is a datetime field
>week_cauculated is a tinyint that store 0 or 1
>
>I first used week_caucluated as bit but I could not create an Index
>
>TIA
>
>
>>I believe that when you return rows through EM, it begins return rows immediately, fetching more rows in the background.
>>
>>Starting with the basics, can you issue a query in QA that returns just one or a few rows? How big is your table. In a C/S environment, you would rarely return all rows form a tx table.
>>
>>Understand that when you return all rows, it is going to be a full table-scan. i.e., no optimization and no indexs (unless I suppose the fields you are returning are all found in an index).
>>
>>FWIW, based on your facts, while there is likely a "problem", it does not rest with SQL Server.
>>
>>
>>>Hi Guys,
>>>
>>>I am having a very strange problem. When I run a query in enterprise manager(server machine,open table, return all rows, click on sql button and type the query there) it runs super fast. When i run the same query in the query analyzer on the server machine it hangs. This also happens If i run this query from a VFp Client or .net. it just hangs. can someone help me.
>>>
>>>The Sql server is running on Windows 2003, 1 gb mem, 300 gb hdd.
>>>The LDF file is 110 GB
>>>the MDF File is 17 GB. No backup of the transaction log has ever been run. The IT Manager only let me know this 1 day ago.
>>>
>>>TIA.