BUG #15679: Partial HASH index takes too much space

From: PG Bug reporting form <noreply(at)postgresql(dot)org>
To: pgsql-bugs(at)lists(dot)postgresql(dot)org
Cc: gentz(at)futuretap(dot)com
Subject: BUG #15679: Partial HASH index takes too much space
Date: 2019-03-08 20:06:45
Message-ID: 15679-5e1b6086b202047c@postgresql.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

The following bug has been logged on the website:

Bug reference: 15679
Logged by: Ortwin Gentz
Email address: gentz(at)futuretap(dot)com
PostgreSQL version: 10.6
Operating system: Amazon RDS
Description:

I created a partial HASH index for a sparsely populated column:

CREATE INDEX partial_hash ON mytable USING HASH(my_id) WHERE my_ID IS NOT
NULL;

Even though the my_id VARCHAR(255) column is populated (NON NULL) only for a
few hundred records, the index takes 256 MB of space (for a table with > 10m
records). Also, it doesn't make a difference if the index is created as a
partial index ("WHERE my_ID IS NOT NULL") or as a full index.

In contrast to that, a BTREE index differs considerably in space for full
and partial:

CREATE INDEX full_btree ON mytable (my_id); # 543 MB
CREATE INDEX partial_btree ON mytable (my_id) WHERE my_ID IS NOT NULL; # 16
KB

See also this StackExchange conversation where user jjanes considers the
behavior a bug in the hash index code:
https://dba.stackexchange.com/a/231660/25337

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message PG Bug reporting form 2019-03-08 21:01:34 BUG #15680: New Versions of pgadmin4 and psqlodbc come with OLD Dlls
Previous Message Andrew Gierth 2019-03-08 19:15:43 Re: BUG #15653: pg_detoast_datum_packed problem