On Tue, Sep 10, 2013 at 5:51 AM, Oliver Kohll - Mailing Lists
<oliver.lists@gtwm.co.uk> wrote:
> On 9 Sep 2013, at 21:03, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:
>
> select string_agg(case when words like '*%*' then upper(btrim(words, '*'))
> else words end, ' ')
> from regexp_split_to_table('The *quick* *brown* fox jumped over the *lazy*
> dog', ' ') as words;
>
> string_agg
> ----------------------------------------------
> The QUICK BROWN fox jumped over the LAZY dog
>
>
> That's quite elegant. In the end I exported and used PERL, as some of my
> 'words' had spaces (they were ingredients like monosodium glutamate), but
> you could probably do a more complex regex in regexp_split_to_table to cope
> with that, or use pl/perl as previously suggested.
IMO, pl/perl is the way to go. Being able to use postgres functions
to transform matched regex expressions would be just wonderful
although I wonder how fast it would be or if it's even possible.
merlin
