http://www.postgresql.org/docs/9.2/static/xfunc-volatility.h...
This allows the planner to optimize queries that use the functions correctly, instead of treating the functions like a black box.
Edit: STABLE would indeed be better, forgot about timezones.
psql=$ \df+ date_trunc
Schema | Name | Result data type | Argument data types | Type | Volatility | Owner | Language | Source code | Description
------------+------------+-----------------------------+-----------------------------------+--------+------------+----------+----------+-------------------+------------------------------------------------------
pg_catalog | date_trunc | interval | text, interval | normal | immutable | postgres | internal | interval_trunc | truncate interval to specified units
pg_catalog | date_trunc | timestamp without time zone | text, timestamp without time zone | normal | immutable | postgres | internal | timestamp_trunc | truncate timestamp to specified units
pg_catalog | date_trunc | timestamp with time zone | text, timestamp with time zone | normal | stable | postgres | internal | timestamptz_trunc | truncate timestamp with time zone to specified unitsFor example:
CREATE OR REPLACE FUNCTION gd_day(timestamptz, text)
RETURNS timestamptz AS
$$
SELECT DATE_TRUNC('day', $1 AT TIME ZONE $2) AT TIME ZONE $2;
$$
LANGUAGE SQL;