"Ozer, Pam" <pozer@automotive.com> writes:
> Isn't this the English standard for collation? Or is this a non-c
> locale as mentioned below? Is there anyway around this?
> LC_COLLATE = 'en_US.utf8'
en_US is probably using somebody's idea of "dictionary order", which
I believe includes ignoring spaces in the first pass. You might be
happier using "C" collation. Unfortunately that requires re-initdb'ing
your database (as of existing PG releases).
regards, tom lane