The "improved" query syntax does reduces the query's time by about 25%.

It's the still unexplained three orders of magnitude difference (23.6 seconds verses 0.015 seconds) that is the limiting factor here. 25% off of 23.6 seconds is, of course, insufficient for a workaround.

Is it your understanding that tags, keys, and collation sequences, alone, should dictate the showplan results of two similarly sized .DBCs?

Next step: My manager has asked me to make a copy of the fast "development" .DBC, zap the tables, import the "production" data, and then see what happens. I will post the results.

Thank you (all) for your kind assistance.

Spencer