Re: Identifying which column matches a full text search

Ryan Wallace wrote:
>
> UPDATE pgweb SET textsearchable_index_col =
> to_tsvector('english', coalesce(title,'') || coalesce(body,''));

> WHERE textsearchable_index_col @@ to_tsquery('create & table')

> Using this approach. Is there any way of retrieving which of the original
> two columns the match was found in?

Afraid not - you're not indexing two columns, you're indexing one:
textsearchable_index_col.

You can add up to four weights to a tsvector though, typically for
title/body matching. See chapter 12.3 for details.

Failing that, where I've had many (a dozen) different sources but want
to search them all I've built a textsearch_blocks table with columns to
identify the source and have triggers that keep it up to date.

--
Richard Huxton
Archonet Ltd