At first I thought the idea of putting all the words into one giant table with an identifier for which subset they came out of was a good idea. Then I tried it.

Basically, MySQL suffers a *huge* performance hit with the large table (600K+ words) vs. the smaller table (350+K words); I think I ran up against some sort of limit in the software. For comparison's sake, though, identical queries took 3.31 seconds for the large table, and 0.76 seconds for the smaller one -- over 4 times as long.

Since it's not possible -- or, as was well pointed out, practical -- to leave the table name as a variable, and since upon further review I only have 4 tables to deal with (I thought it was going to be more like 12, but I ruled a bunch of word subsets out), I'm just going to have 4 separate statements prepped up (and yes, one database handle *can* support multiple queries -- I've done it before).

Thanks for your help anyway.

Alex Kirk