Efficient way to count rows when using GROUP BY

Hi All, I use SELECT COUNT(*) FROM whatever WHERE whatever to determine how many rows a particular query will return (then use this count to assist with pagination). The problem is that this does not work if the query contains a GROUP BY statement as, of course, the query will return many rows each with a count by each grouping. Does anyone know an efficient way to determine the number of ROWS a query with a GROUP BY statement will return (without just executing the entire query with no LIMIT statement)? Thanks ... Ross