"Joshua D. Drake" <jd@commandprompt.com> writes:
>> It's fairly clear that we could support concurrent builds of nonunique
>> indexes, but is that enough of a use-case to justify it?
> I believe there would be. Most PostgreSQL users I run into, develop in
> production, which means being able to add an index they forgot when
> doing query analysis.
True, unique constraints are usually something you should get right to
start with. But it'll be annoying if we can do everything BUT that :-(
regards, tom lane