Tom Lane-2 wrote
> David G Johnston <
> david.g.johnston@
> > writes:
>> Suggest wrapping it in a function - or making a "regexp_matches_single"
>> function that behaves similarly but returns a single text[] instead of a
>> SETOF text[]
>
> I wonder if we should have such a thing built-in. This isn't the first
> complaint we've heard about the SETOF API being awkward to use, and it's
> only needed if you specify the 'g' flag. Perhaps we could have a variant
> that forbids 'g' and returns a non-set (either the single match, or NULL).
>
> regards, tom lane
+1
Given that regexp_matches is in core having this most useful API makes
perfect sense to me. Parsing out components from data is an extremely
common use-case for regexp - one that I personally encounter much more often
than needing global matching behavior.
I agree that the presence of the 'g' flag when calling regexp_matches_single
(_once?) should emit an error.
David J.
--
View this message in context:
http://postgresql.1045698.n5.nabble.com/BUG-10889-Cannot-add-2-floats-from-regular-expression-tp5810748p5810756.html
Sent from the PostgreSQL - bugs mailing list archive at Nabble.com.