On 03.03.2017 16:17, Nicolas Paris wrote:
> Nice ! I do have 9.6 version.
>
> Would this kind of index could handle more than 20M large texts ? The
> recheck condition looks ressource consuming.
You are right. I think pg_trgm will be not good for such large texts,
unfortunately.
>
> The full text index + phrase search + synonym dictionnary is the only
> other alternativ to deal with typo-phrase mining ?
I suppose there are no other options now. Though, prefix search maybe
will help you [1].
>
> Is there any possibility in the future to add typo in the full text
> road-map ?
As far as I know, there is no plans in the near future to add similarity
full text search.
1.
https://www.postgresql.org/docs/current/static/textsearch-dictionaries.html
--
Artur Zakirov
Postgres Professional: http://www.postgrespro.com
Russian Postgres Company