Re: BUG #3235: Partitioning has problem with timestamp and timestamptz data types

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "Christian Gonzalez" <christian(dot)gonzalez(at)sigis(dot)com(dot)ve>
Cc: pgsql-bugs(at)postgresql(dot)org
Subject: Re: BUG #3235: Partitioning has problem with timestamp and timestamptz data types
Date: 2007-04-18 05:24:46
Message-ID: 20938.1176873886@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

"Christian Gonzalez" <christian(dot)gonzalez(at)sigis(dot)com(dot)ve> writes:
> Description: Partitioning has problem with timestamp and timestamptz
> data types

In your example as given, you partition the table on "logdate", and then
add an unrelated column "logdatet" and seem to expect that the system
will think that's a partitioning condition.

I think you probably misstated your example, and meant to complain that
a table properly partitioned on a timestamptz column isn't behaving as
you'd wish for comparisons against "date" values. The problem here is
that the conversion is dependent on TimeZone, so is not immutable, so
the planner daren't depend on it to prove that the unwanted partition
tables need not be searched. You need to think harder about how
timezones should enter into your search rules...

regards, tom lane

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Tom Lane 2007-04-18 05:28:05 Re: BUG #3232: Regression: pgsql server startup problem with encrypted partitions
Previous Message Robert wang 2007-04-18 03:10:41 BUG #3237: function to_char() returns wrong value