Andres, Tom,
> > I suspect that we'd first need a patch to refactor the existing copy code a
> > good bit to clean things up. After that it hopefully will be possible to plug
> > in a new format without being too intrusive.
>
> I think that step 1 ought to be to convert the existing formats into
> plug-ins, and demonstrate that there's no significant loss of performance.
Yep, this looks like a promising strategy to me too.
> I know little about parquet - can it support FROM STDIN efficiently?
Parquet is a compressed binary format with data grouped by columns
[1]. I wouldn't assume that this is a primary use case for this
particular format.
[1]: https://parquet.apache.org/docs/file-format/
--
Best regards,
Aleksander Alekseev