On 02/01/2008, Matthew Wilson <matt@tplus1.com> wrote:
> On Wed 02 Jan 2008 04:23:46 AM EST, Pavel Stehule wrote:
> > Hello
> >
> > IDEA 3:
> >
> > Use two hash functions:
> >
> > CREATE UNIQUE INDEX uidx ON TEST((decode(md5(a),'hex')),(hashtext(a)));
> >
> > removing spaces helps
> > CREATE UNIQUE INDEX uidx ON test((decode(md5(lower(replace(a,' ',''))),'hex')));
> >
> > Regards
> > Pavel Stehule
> >
> > CREATE UNIQUE INDEX nodups on MESSAGE (my_timestamp_col, md5(my_text_col));
>
> What is the advantage of this method?
>
much less unique index.
Regards
Pavel
> --
> Programming, economics, gardening, life in Cleveland.
> http://blog.tplus1.com
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Have you searched our list archives?
>
> http://archives.postgresql.org/
>