Cursors In SQL Server

I have read ample technical papers on Cursors.  I like to use cursors because it gives you a row by row control in a stored procedure.  But most articles I have read advise against the use of cursors (server side).I typically use read only or Forward Only Cursors.  This seems to be the best way to transpose data.My questions is, is there any guildlines to using cursors?  Do you look at the performance of the cursor?  If the query runs in 30 seconds and the cursor process is 15 seconds, is that acceptable (based on a million transactions, but filtering 1000-2000 records based on indexed fields.)?Or is it better to break cursor based steps  into multiple queries with insert / update to generate a transposed output? Even if this involves a base query with multiple joins. Is it better to hammer SQL Server once and then output the results on a row by row, column by column  basis, or do multiple queries.The bottom line is I have been able to provide the required output using cursors, but I just don't want to potentially bog down the system.Can anyone list specific cases where a cursor was a better method than multiple queries?  And particularily with large volume data mining operations. Thanks DotNetNow