From: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>, Stephan Szabo <sszabo(at)megazone(dot)bigpanda(dot)com>, Andrew Dunstan <andrew(at)dunslane(dot)net>, "Jim C(dot) Nasby" <jnasby(at)pervasive(dot)com>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Reducing the overhead of NUMERIC data |
Date: | 2005-11-06 08:45:58 |
Message-ID: | 1131266758.8300.2051.camel@localhost.localdomain |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers pgsql-patches |
On Thu, 2005-11-03 at 10:32 -0500, Tom Lane wrote:
> I'd feel a lot happier about this if we could keep the dynamic range
> up to, say, 10^512 so that it's still true that NUMERIC can be a
> universal parse-time representation. That would also make it even
> more unlikely that anyone would complain about loss of functionality.
>
> To do that we'd need 8 bits for weight (-128..127 for a base-10K
> exponent is enough) but we need 9 bits for dscale which does not
> quite fit. I think we could make it go by cramming the sign and
> the high-order dscale bit into the first NumericDigit --- the
> digit itself can only be 0..9999 so there are a couple of bits
> to spare. This probably *would* slow down packing and unpacking of
> numerics, but just by a couple lines of C. Arguably the net reduction
> in I/O costs would justify that.
I've got a working version of the code using the above scheme, with
these additional wrinkles:
NaN is indicated by weight=-128, giving a dynamic range of 10^508.
Zeroes are fully compressed, except when the Scale > 255. In that case,
the first digit is present to signify the presence of the high order
Scale bit.
Comments?
Once 8.1 is released, I'll go back and see if I can improve the coding
in a few days with fresh eyes, then submit a patch.
Best Regards, Simon Riggs
From | Date | Subject | |
---|---|---|---|
Next Message | Simon Riggs | 2005-11-06 09:00:05 | Re: [HACKERS] insert performance for win32 |
Previous Message | Tom Lane | 2005-11-05 21:47:41 | Last chance to defend RTREE index access method |
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2005-11-06 16:26:39 | Re: Reducing the overhead of NUMERIC data |
Previous Message | Harald Fuchs | 2005-11-05 15:34:56 | Re: Reducing the overhead of NUMERIC data |