You are viewing a plain text version of this content. The canonical link for it is here.
Posted to derby-dev@db.apache.org by Mike Matrigali <mi...@sbcglobal.net> on 2007/05/14 21:57:06 UTC
is there any territory based collation that is exactly the same as
the default binary based collation?
I was hoping once the collation work is mostly done to run a full set
of nightly tests somehow setting all the databases to be collation based
on some territory - hoping a huge percentage would work (ie. those not
dependent on a
specific territory). The best case would be a territory with ordering
that exactly matches the binary default of string comparison, but my
guess is a huge percentage of tests would be ok with a territory that
just handled same collation of ascii values (or maybe even just 26
lower, 26 upper and numbers).
At first I thought english might work, but ordering of cap vs lower case
is different. So I was wondering if anyone knew an ordering that
would be better than english for data that is probably almost all
ascii characters.
I also don't quite know how I was going to do this. Worst case I
thought I might just hack in my own codeline the create database
code to recognize a default setting of territory/collation and in
that case override it to something. Seemed like this is a problem
that junit suites were setup to handle so maybe there is something
more elegant I could do there?