RE: JOIN databases of different charsets

> From: Fernando Nunes [mailto:domusonline@gmail.com] > > I believe Ian is missing the point. The issue is not unloading data and > load it into a different locale. > Apparently the issue is that the applications require distributed > queries between two (or more) databases with different locales. You are correct. > I believe this could be achieved with an undocumented variable setup in > the environment. > Check this article for further information: > > https://www.ibm.com/developerworks/mydeveloperworks/blogs/gbowerman/ent > ry/trouble_with_locales?lang=en > > Now... The idea behind this variable was not allowing this. But the > side effect is that it will... The question is: Will this be > acceptable? > I believe you may hit several problems. You may want to talk to support > about this, but I suppose the answer will be that it's unsupported and > can potentially lead to corruption or bad query results... I'll check this out on a test instance. See what the queries with JOINS will return, I doubt the results will be good. As long as we're doing SELECT/read-only then there should not be any corruption though? Then again the idea of a misconfigured client screwing up the tables with incorrect locales sounds bad enough. The correct solution is probably to push the migrations and get this over with as fast as possible. The sooner we're in utf only land, the better. Regards, Aleksander Kamenik System Administrator Krediidiinfo AS an Experian Company Phone: +372 665 9649 Email: aleksander@krediidiinfo.ee