Re: Analysis Function - Mailing list pgsql-performance

From Tom Lane
Subject Re: Analysis Function
Date
Msg-id 13998.1276443260@sss.pgh.pa.us
Whole thread Raw
In response to Re: Analysis Function  (Heikki Linnakangas <heikki.linnakangas@enterprisedb.com>)
Responses Re: Analysis Function
List pgsql-performance
Heikki Linnakangas <heikki.linnakangas@enterprisedb.com> writes:
> We could have a function like:

> construct_timestamp(year int4, month int4, date int4, hour int4, minute
> int4, second int4, milliseconds int4, timezone text)

This fails to allow specification to the microsecond level (and note
that with float timestamps even smaller fractions have potential use).
I would suggest dropping the milliseconds argument and instead letting
the seconds arg be float8.  That seems a closer match to the way people
think about the textual representation.

> Now that we have named parameter notation, callers can use it to
> conveniently fill in only the fields needed:

It's not immediately obvious what the default value of "timezone"
will be?

            regards, tom lane

pgsql-performance by date:

Previous
From: Anj Adu
Date:
Subject: Re: out of memory
Next
From: Tom Lane
Date:
Subject: Re: Analysis Function