Rick Schumeyer wrote:
> I want to be able to search a list of articles for title words as well
> as author names.... I'm not sure the best strategy for the names. The
> full text parser "parses" the names giving undesirable results.
>
> For example,
>
> select to_tsvector('claude Jones');
> to_tsvector
> --------------------
> 'jone':2 'claud':1
>
>
> Is there a way to tell the parser to index the words in a column without
> trying to determine word roots?
> Or what is the best way to index names for efficient searching?
I've got a similar question; but would love a dictionary that
could give me an efficient index that considers that
Bill/William and Bob/Rob/Robert and Khadaffi/Qaddafi might
share the same roots. Ideally it'd return exact matches
first, followed by the similar terms.
I kludged up some many step queries to try to do this; but
wonder if this would work better as a tsearch dictionary,
and wonder even more if I'm re-inventing something that's
already out there.
Anyone know of such a tsearch dictionary that is aware of
the roots of names?