Roger <[EMAIL PROTECTED]> writes: > Have you tried creating indexes on your rows. > [..] > I suggest you add indexes on text_val
Yes. I use create index text_val_idx on geodb_textdata(text_val); I just experimented with dropping/recreating indexes and it sometimes even feels a bit faster *without* the index. Anyway, both variants are slow, nearly same speed. Another idea: the data are utf-8, can this be a problem? Do I have to declare this somewhere at import ore runtime? Can I set it to non-unicode, to see performance difference? (Greeti+Tha)nX Steffen -- Steffen Schwigon <[EMAIL PROTECTED]> Dresden Perl Mongers <http://dresden-pm.org/>