I have an indexed column where a simple "SELECT MAX(<column>) FROM
<table>" generates a full table scan. I guess it has to do with the
index compresion stuff I read somewhere. This particular table has
over 1.5 million records (the database has plenty of super populated
tables).

Should I create an ascending AND a descending index to obtain decent
perfomance? What about storage... will it duplicate?

BTW, how does index compression affect negatively on geographic
coordinates (lat/lon, UTM)? If there is a problem, is there any
workaround (perhaps storing hours, minutes, seconds instead of
decimal notation?).

TIA.

Greeting,
Daniel Berstein