RE: [BUG]Missing REPLICA IDENTITY check when DROP NOT NULL

On Wed, Nov 24, 2021 7:29 PM, Michael Paquier <michael@paquier.xyz> wrote:
> 
> On Wed, Nov 24, 2021 at 07:04:51AM +0000, tanghy.fnst@fujitsu.com wrote:
> > create table tbl (a int not null unique);
> > alter table tbl replica identity using INDEX tbl_a_key;
> > alter table tbl alter column a drop not null;
> > insert into tbl values (null);
> 
> Oops.  Yes, that's obviously not good.
> 
> > To solve the above problem, I think it's better to add a check when
> > executing  ALTER COLUMN DROP NOT NULL,
> > and report an error if this column is part of replica identity.
> 
> I'd say that you are right to block the operation.  I'll try to play a
> bit with this stuff tomorrow.
> 
> > Attaching a patch that disallow DROP NOT NULL on a column if it's in
> > a REPLICA IDENTITY index. Also added a test in it.
> 
>        if (indexStruct->indkey.values[i] == attnum)
>            ereport(ERROR,
>                    (errcode(ERRCODE_INVALID_TABLE_DEFINITION),
> -                   errmsg("column \"%s\" is in a primary key",
> +                   errmsg(ngettext("column \"%s\" is in a primary key",
> +                                   "column \"%s\" is in a REPLICA IDENTITY index",
> +                                   indexStruct->indisprimary),
>                     colName)));
> Using ngettext() looks incorrect to me here as it is used to get the
> plural form of a string, so you'd better make these completely
> separated instead.

Thanks for your comment. I agree with you.
I have fixed it and attached v2 patch.

Regards,
Tang