On Sat, Jul 17, 2010 at 07:46:23PM +0200, Thomas Kellerer wrote:
> Tom Lane wrote on 17.07.2010 19:35:
>> I'd dismiss those numbers as being within experimental
>> error, except it seems odd that they all differ in the same direction.
>
> And it's reproducable (at least on my computer). As I said I
Tom Lane wrote on 17.07.2010 19:35:
Thomas Kellerer writes:
Tom Lane wrote on 17.07.2010 16:36:
Well, nobody's offered any actual *numbers* here.
I measured the runtime as seen from the JDBC client and as reported by explain analyze
(the last line reading "Total runtime:")
The "runtime"
Thomas Kellerer writes:
> Tom Lane wrote on 17.07.2010 16:36:
>> Well, nobody's offered any actual *numbers* here.
> I measured the runtime as seen from the JDBC client and as reported by
> explain analyze (the last line reading "Total runtime:")
The "runtime" from explain analyze really should
Thomas Kellerer wrote on 17.07.2010 18:29:
Want to do some experiments?
Apparently there *is* a substiantial overhead, but I suspected the
sending of the raw SQL literal to be a major factor here.
(Server and JDBC program were running on the same machine)
In case any one is interested.
Out
Tom Lane wrote on 17.07.2010 16:36:
Thomas Kellerer writes:
I'm till a bit surprised that parsing the statement _with_ a column list is
mesurably slower than withou a column list.
Well, nobody's offered any actual *numbers* here. It's clear that
parsing the column list will take more time t
Thomas Kellerer writes:
> I'm till a bit surprised that parsing the statement _with_ a column list is
> mesurably slower than withou a column list.
Well, nobody's offered any actual *numbers* here. It's clear that
parsing the column list will take more time than not doing so, but
whether that a
Craig Ringer wrote on 17.07.2010 03:13:
On 17/07/10 04:26, Thomas Kellerer wrote:
Hmm.
For years I have been advocating to always use fully qualified column
lists in INSERTs (for clarity and stability)
And now I learn it's slower when I do so :(
If you're not doing hundreds of thousands of id
On 17/07/10 04:26, Thomas Kellerer wrote:
> Hmm.
> For years I have been advocating to always use fully qualified column
> lists in INSERTs (for clarity and stability)
> And now I learn it's slower when I do so :(
If you're not doing hundreds of thousands of identical ones at a time,
it's still v
Tom Lane wrote on 16.07.2010 18:40:
Thomas Kellerer writes:
the explanation of the --inserts option of pg_dumps states that
"The --column-inserts option is safe against column order changes, though even
slower."
The way I read this is, that
INSERT INTO table (column, ...) VALUES ...
Thomas Kellerer writes:
> the explanation of the --inserts option of pg_dumps states that
> "The --column-inserts option is safe against column order changes, though
> even slower."
> The way I read this is, that
>INSERT INTO table (column, ...) VALUES ...
> is slower than
>INSERT INTO
Hi,
the explanation of the --inserts option of pg_dumps states that
"The --column-inserts option is safe against column order changes, though even
slower."
The way I read this is, that
INSERT INTO table (column, ...) VALUES ...
is slower than
INSERT INTO table VALUES ...
Is that really tr
11 matches
Mail list logo
