Subject RE: [firebird-support] duplicates in a table with PK
Author Puigsegur, Jordi
Hi Helen,

> You already did - gstat -h told you the default character set is
> NONE. That means data are stored in straight ascii format. The
> database won't know how to locate strings containing accented
> characters. Your sample data, at least, doesn't appear to contain
> accented characters.

Data in the VARCHAR(100) should not contain accented characters, but it might contain Ñ. Don't know if that is a problem.

> If you are storing strings with Spanish accented characters, using
> character set NONE is going to bite you one way or another, since it
> recognises only the US ASCII set of characters as characters.
> I suggest you treat today's database rebuild as a practice exercise
> but rebuild the for-production database with a Spanish-aware
> character set as the default character set and a suitable collation
> on indexed fields to allow searching and matching to make
> sense. Then make sure that your client programs connect with that
> character set.

Ok. We'll do that. For new databases is it enough to select CHARSET ISO8859_1 in FlameRobin's "Create new database" window? Will that set the correct collation as well? How can I rebuild the old one with the correct settings. Using gbak maybe?

Finally, any idea on why the duplicate values appeared ?

Thanks for the advise.