Re: Add support for DEFAULT specification in COPY FROM

On Mon, Sep 26, 2022 at 8:12 AM Israel Barth Rubio <barthisrael(at)gmail(dot)com>
wrote:

> Hello Andrew,
>
> > . There needs to be a check that this is being used with COPY FROM, and
> > the restriction needs to be stated in the docs and tested for. c.f.
> > FORCE NULL.
> >
> > . There needs to be support for this in psql's tab_complete.c, and
> > appropriate tests added
> >
> > . There needs to be support for it in contrib/file_fdw/file_fdw.c, and a
> > test added
> >
> > . The tests should include psql's \copy as well as sql COPY
> >
> > . I'm not sure we need a separate regression test file for this.
> > Probably these tests can go at the end of src/test/regress/sql/copy2.sql.
>
> Thanks for your review! I have applied the suggested changes, and I'm
> submitting the new patch version.
>
> Kind regards,
> Israel.
>

Hi,

+ /* attribute is NOT to be copied from input */

I think saying `is NOT copied from input` should suffice.

+ defaults = (bool *) palloc0(num_phys_attrs * sizeof(bool));
+ MemSet(defaults, false, num_phys_attrs * sizeof(bool));

Is the MemSet() call necessary ?

+ /* fieldno is 0-index and attnum is 1-index */

0-index -> 0-indexed

Cheers