On Fri, Jan 11, 2013 at 9:02 AM, Heikki Linnakangas <hlinnakangas@xxxxxxxxxx> wrote:
On 11.01.2013 12:36, Matheus de Oliveira wrote:
One thing that caught my eye:
CREATE OR REPLACE FUNCTION partition_insert_trigger_spi()
RETURNS triggerAS 'partition_insert_trigger_spi'**,'partition_insert_trigger_**spi'
LANGUAGE C
VOLATILE STRICT
SET DateStyle TO 'ISO';
Calling a function with SET options has a fair amount of overhead, to
set/restore the GUC on every invocation. That should be avoided in a
performance critical function like this.
I (stupidly) used SPI_getvalue [1] and expected it to always return as
YYYY-MM-DD, but them I remembered it would do that only with DateStyle=ISO.
But the truth is that I couldn't see any overhead, because the function was
without that on my first tests, and after that I saw no difference on the
tests.
Oh, ok then. I would've expected it to make a measurable difference.The fastest way is probably to use j2date like date_out does:
I think I should use SPI_getbinvalue instead, but I don't know how
to parse the result to get year and month, any help on that?
DateADT date = DatumGetDateADT(x)
int year, month, mday;
if (DATE_NOT_FINITE(date))
elog(ERROR, "date must be finite");
j2date(date + POSTGRES_EPOCH_JDATE, &year, &month, &mday);
- Heikki
Nice. With the modifications you suggested I did saw a good improvement on the function using SPI (and a little one with heap_insert). So I was wrong to think that change the GUC would not make to much difference, the SPI code now runs almost as fast as the heap_insert:
heap_insert: 31896.098 ms
SPI: 36558.564
Of course I still could make some improvements on it, like using a LRU to keep the plans, or something like that.
The new code is at github.
Regards,
--
Matheus de Oliveira
Analista de Banco de Dados
Dextra Sistemas - MPS.Br nível F!
www.dextra.com.br/postgres
