Hi Matteo,
2013/7/2 Matteo Beccati
> I'm talking about parameters, not results. The speedup would be only for
> inserts, not selects (which would require something that's not available
> in libpq).
>
For insert or update, binary can be inserted/updated directly.
If we do this transparently to u
On 02/07/2013 12:52, Yasuo Ohgaki wrote:
> Hi Matteo,
>
> 2013/7/2 Matteo Beccati mailto:p...@beccati.com>>
>>
>> > The API is not for ease of use, but for better performance as it
> eliminate
>> > needless hex conversions on both server and client side.
>>
>> What I meant is that we might have th
Hi Matteo,
2013/7/2 Matteo Beccati
>
> > The API is not for ease of use, but for better performance as it
eliminate
> > needless hex conversions on both server and client side.
>
> What I meant is that we might have that speed improvements for free in
> PDO when bytea is used as a parameter (i.e.
On 02/07/2013 12:21, Yasuo Ohgaki wrote:
> 2013/7/2 Matteo Beccati mailto:p...@beccati.com>>
> We could easily (and transparently) use binary for the parameters
> explicitly bound as LOB, but I see no way we can use binary for results
> -- which usually is the most common operation. May
Hi Matteo,
2013/7/2 Matteo Beccati
> > AFAIK, PDO doesn't handle bytea data natively.
>
> It does, as far as I remember. See PDO::PARAM_LOB
>
> We could easily (and transparently) use binary for the parameters
> explicitly bound as LOB, but I see no way we can use binary for results
> -- which
On 26/06/2013 11:51, Yasuo Ohgaki wrote:
> Hi Matteo,
>
> PDO_pgsql on the other hand, when used with explicit parameter binding,
> could automatically use binary format for LOB fields.
>
> AFAIK, PDO doesn't handle bytea data natively.
It does, as far as I remember. See PDO::PARAM_LOB
Hi Matteo,
2013/6/26 Matteo Beccati
> I have mixed feelings... I see where this might come in handy, but I
> think it would be a bit too user-unfriendly.
>
I have mixed feelings for supporting direct binary support, too.
Main intension is faster bytea(binary) handling.
Without binary support,
On 26/06/2013 00:44, Yasuo Ohgaki wrote:
> New:
> resource pg_execute ([ resource $connection ], string $stmtname , array
> $params [, array $params_format [, bool $binary_result]] )
>
> Any thoughts?
Using binary format "requires knowledge of the internal representation
expected by the backend.
Sorry, missed the list...
On 26 June 2013 09:54, Yasuo Ohgaki wrote:
> Hi Make,
>
> 2013/6/26 Michael Wallner
>>
>> I didn't look at the code yet, but how do you know about binary format
>> conventions of all possible types returned?
>
> Users can only specify if the result is returned as text o
Hi all,
Current pgsql module lacks direct binary support. Therefore, users have to
use pg_escape_bytea/pg_unescape_bytea for binary handling even with
prepared statement type APIs. Use of
pg_escape_bytea/pg_unescape_bytea requires some overheads on both server
and client side.
Following patch ena
10 matches
Mail list logo
