Re: Mixing greediness in regexp_matches

Tom Lane wrote:

> regression=# select regexp_split_to_array('junkfoolbarfoolishfoobarmore',
> 'foo|bar|foobar');
> regexp_split_to_array
> -----------------------
> {junk,l,"",lish,more}
> (1 row)
>
> The idea would be to iterate over the array elements, tracking the
> corresponding position in the source string, and re-discovering at
> each break which of the original alternatives must've matched.
>
> It's sort of annoying that we don't have a simple "regexp_location"
> function that would give you back the starting position of the
> first match.

It occurred to me too that regexp_split_to_table or array would make
this problem really easy if only it had a mode to capture and return the
matched parts too.

FWIW, in plperl, there's a simple solution:

$string =~ s/(foobar|foo|...)/$replace{$1}/g

when %replace is a hash of the substitutions %(foo=>baz,...).
The strings in the alternation are tested in their order of
appearance, so you can choose to be greedy or not by just sorting
them by length.

Best regards,
--
Daniel Vérité
PostgreSQL-powered mailer: http://www.manitou-mail.org
Twitter: @DanielVerite