Re: [psycopg] speed concerns with executemany()

On 12/30/2016 06:42 PM, Christophe Pettus wrote:
>
>> On Dec 30, 2016, at 14:24, Daniele Varrazzo <daniele.varrazzo@gmail.com> wrote:
>>
>> The implementation of executemany as described by me a few days ago is
>> available in this gist, not heavily tested:
>>
>> https://gist.github.com/dvarrazzo/4204cca5d1cd7c9c95bc814d81e6b83e
>>
>> I would like to know if anyone sees any shortcoming in this new implementation.
>
> Seems fine to me!  I wish there was a way of feeding the queries down the pipe asynchronously, rather than having to
builda gigantic string, but that's probably more trouble than the feature is worth. 

you'd need to be careful with that as a series of multiple parameter
sets may have dependencies on each other, not to mention people might
find it surprising that sequences / defaults / SERIAL etc. aren't firing
off in the order in which parameter sets were given.




>
> --
> -- Christophe Pettus
>    xof@thebuild.com
>