| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Max Fomichev <max(dot)fomitchev(at)gmail(dot)com> |
| Cc: | pgsql-novice(at)lists(dot)postgresql(dot)org |
| Subject: | Re: GIN, pg_trgm and large table |
| Date: | 2018-03-10 17:00:27 |
| Message-ID: | 22627.1520701227@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-novice |
Max Fomichev <max(dot)fomitchev(at)gmail(dot)com> writes:
> I have the following table and index with about 15 billion records.
> ...
> PostgreSQL 9.5.12 on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu
> 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609, 64-bit
> ...
> It there any way to improve GIN/pg_tgrm performance on a such large table?
There was some work done in 9.6 to improve pg_trgm's performance when
dealing with queries involving very common trigrams. So maybe an update
to 9.6 or v10 would help you.
I have a bad feeling though that 15 billion rows is too many for an index
based on trigrams to be really useful --- there are just not enough
distinct trigrams. It's too bad we don't have a more general N-gram
indexing facility.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Max Fomichev | 2018-03-10 17:44:21 | Re: GIN, pg_trgm and large table |
| Previous Message | Max Fomichev | 2018-03-10 10:31:46 | GIN, pg_trgm and large table |