Postgres Pro
Downloads
Home   >   mailing lists

Re: Which is faster: md5() or hashtext()? - Mailing list pgsql-general

From Grzegorz Jaśkiewicz
Subject Re: Which is faster: md5() or hashtext()?
Date
Msg-id AANLkTimM0HtHrmujgngNPSmGFZPwnHE+XSwqSr6N5-4M@mail.gmail.com
Whole thread Raw
In response to Which is faster: md5() or hashtext()?  ("Henry C." <henka@cityweb.co.za>)
Responses Re: Which is faster: md5() or hashtext()?  ("Henry C." <henka@cityweb.co.za>)
List pgsql-general
Tree view 
Timing is on.
psql (9.1devel)
Type "help" for help.

# select count(hashtext(a::text)) FROM generate_series(1,10000) a;
 count
-------
 10000
(1 row)

Time: 106.637 ms
# select count(hashtext(a::text)) FROM generate_series(1,1000000) a;
  count
---------
 1000000
(1 row)

Time: 770.823 ms
# select count(md5(a::text)) FROM generate_series(1,1000000) a;
  count
---------
 1000000
(1 row)

Time: 1238.453 ms
# select count(hashtext(a::text)) FROM generate_series(1,1000000) a;
  count
---------
 1000000
(1 row)

Time: 763.169 ms
# select count(md5(a::text)) FROM generate_series(1,1000000) a;
  count
---------
 1000000
(1 row)

Time: 1258.958 ms


I would say hashtext is consequently beating md5 in terms of performance here.

Just remember, that it returns integer, unlike md5 that returns text.

pgsql-general by date:

Previous
From: "Henry C."
Date:
Subject: Which is faster: md5() or hashtext()?
Next
From: "Andrus"
Date:
Subject: Cloning database without dump/restore