2006/12/18, Jeff Davis <pgsql@j-davis.com>:
> On Fri, 2006-12-15 at 17:21 -0200, Clodoaldo wrote:
> > -- drop table test_table;
> > create table test_table (tname varchar, value integer);
> > insert into test_table values ('[ab]x', 1);
> > insert into test_table values ('[ab]y', 2);
> > insert into test_table values ('[Ab]z', 3);
> > insert into test_table values ('w[aB]', 8);
> > insert into test_table values ('[abx', 4);
> > insert into test_table values ('ab]x', 5);
> > insert into test_table values ('xyz', 6);
> > insert into test_table values ('Xyz', 7);
> >
> ...
> > total | tname | value_total
> > -------+-------+-------------
> > 4 | [AB] | 14
> > 1 | [abx | 4
> > 1 | ab]x | 5
> > 1 | xyz | 6
> > 1 | Xyz | 7
> > (5 rows)
> >
>
> It looks like what you need is a function that returns the upper() of
> the substring enclosed by "[]" if that substring exists, or else the
> original string.
>
> In other words, make a user-defined function that turns the string into
> "tname" and GROUP BY the result of that function.
>
That worked. I was fixed in solving it with a regular expression and
didn't see the obvious.
Thanks.
--
Clodoaldo Pinto Neto