Re: index unique - Mailing list pgsql-general

From Peter J. Holzer
Subject Re: index unique
Date
Msg-id 20210608184212.GA9580@hjp.at
Whole thread Raw
In response to Re: index unique  (Marc Millas <marc.millas@mokadb.com>)
Responses Re: index unique  (Marc Millas <marc.millas@mokadb.com>)
List pgsql-general
On 2021-06-08 18:30:16 +0200, Marc Millas wrote:
> the only pb observed is the size of the object accepted. if the geom is a bit
> "big" then the index errors.about btree size of index object.
> but if I create a table test_l with a text column blabla as a PK, and insert a
> 100 000 character long string , no pb.
> if I do an explain analyze select blabla from test_l order by blabla, postgres
> tells me that it did an index only scan.
> so, I wonder why Postgres is able to put a 100 000 long text in a btree index
> and NOT a geom column which wkt is 10 000 bytes long.

Good question. Maybe your texts compresses better than your geometries:

hjp=> insert into t(t) values(repeat('a', 235327) || '1');
INSERT 0 1
Time: 60.057 ms
hjp=> insert into t(t) values(repeat('a', 235328) || '1');
ERROR:  index row size 2720 exceeds maximum 2712 for index "t_pkey1"
HINT:  Values larger than 1/3 of a buffer page cannot be indexed.
Consider a function index of an MD5 hash of the value, or use full text indexing.
Time: 58.751 ms

Note the difference between the length of the string I was trying to
insert and the length of the row it complains about.

        hp

--
   _  | Peter J. Holzer    | Story must make more sense than reality.
|_|_) |                    |
| |   | hjp@hjp.at         |    -- Charles Stross, "Creative writing
__/   | http://www.hjp.at/ |       challenge!"

Attachment

pgsql-general by date:

Previous
From: Marc Millas
Date:
Subject: Re: index unique
Next
From: Marc Millas
Date:
Subject: Re: index unique