This is a bit of a special problem as we are replicating a large existng database that is based on a COBOL type language. the table srtucture is designed to allow us to use existing clients but redirecting to a SQL server. What I don't understand is that:

mysql> select keycase,keydate,nm1fl from Log where nm1fl='BE EASY DOC' order by nm1fl limit 10; +-----------+---------+-------------+ | keycase | keydate | nm1fl | +-----------+---------+-------------+ | QH 100000 | 970106 | BE EASY DOC | +-----------+---------+-------------+ 1 row in set (0.40 sec)

comes back in a respectable time. If the table has a key on the nm1fl column then the next 10 should be able to be found by just getting the next 10 in the index. Instead it does a table scan or something. I'm still not sure why 'having' gives a response and 'where' starts thrashing.

mysql> select keycase,keydate,nm1fl from Log having nm1fl='BE EASY DOC' order by nm1fl limit 10; +-----------+---------+-------------+ | keycase | keydate | nm1fl | +-----------+---------+-------------+ | QH 100000 | 970106 | BE EASY DOC | +-----------+---------+-------------+ 1 row in set (14.06 sec)

and why does 'having' take so much longer with this same query?