Re: FYI: porting Copy API to 8.x

Hi Kalle,


On 30-May-07, at 3:45 PM, Kalle Hallivuori wrote:

> Hi Dave!
>
> 2007/5/30, Dave Cramer <pg@fastcrypt.com>:
>> The reason that the patch has never been officially accepted is
>> because it creates two protocol paths.
>>
>> To do this correctly you need to add the copy functionality into the
>> current protocol handler as opposed to having a separate protocol
>> handler for copy.
>
> Thanks, now I understand why it's not there. I can take a look at
> alternative approaches tomorrow, and only go with the ugly one if I
> don't come up with an (moderately easily implementable) acceptable
> design.
>
> Is it more of an open design problem that many people have looked
> closely into, or rather just a matter of someone getting to it? Ie. is
> it likely there is no clean way to add these special cases inside the
> protocol handler?
>
> As for redesign, which would be the least abhorred way of providing
> the functionality for use:
>
> 1. Separate Copy API available via a getCopyAPI() method in
> PGConnection API?
> 2. Separate Copy API available with some cleaner approach, which?
> 3. Special Copy methods in PGConnection API?
> 4. Special cases to batch processing (FROM STDIN) and resultset
> collecting (TO STDOUT)?
> 5. Special cases to some other part of the standard JDBC API, which?
>
I think you have to make it an extension somehow. I'd leave
getCopyAPI in place, I think the harder part is getting the protocol
to work,
> Cheers,
>
> --
> Kalle Hallivuori +358-41-5053073 http://korpiq.iki.fi/
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 3: Have you checked our extensive FAQ?
>
>               http://www.postgresql.org/docs/faq