RE: JOIN databases of different charsets

> From: Ian Michael Gumby [mailto:im_gumby@hotmail.com] > It looks like you already know the answer to your question. > Different character sets will order/compare differently. Its not like > you're writing your own comparable operator which would allow you to > handle different character sets. Well theoretically it would be possible for informix first convert the ISO results to UTF and then do the JOIN. It's not a very popular feature request I guess. > At first I thought it would be a cool feature if you could specify the > character sets on a per table basis, but then if you think about it... > just go UTF-8 which should handle most if not all of your character set > needs. (I haven't played with Kanji or other character sets.) That's why we're migrating to UTF. But it's not as simple as just a unload/load. > Since you're going to be converting from one character set to > another... you could do it as you unload the table, or as you load the > new database table. The real issue is that we are slowly migrating from one big iso database to several smaller ones in UTF. It's not just a charset change per table. New projects use new UTF databases, but they also need data from the old iso database while not everything has been migrated/updated. New tables with new schemas are created usually. It'll probably take more than a year for everything to migrate to UTF databases. Old projects are rewritten for UTF too, which don't need a schema changes, but it will still take time. Meanwhile not being able to make JOINs between new and old databases sucks a bit. Thanks for confirming, Aleksander Kamenik System Administrator Krediidiinfo AS an Experian Company Phone: +372 665 9649 Email: aleksander@krediidiinfo.ee