On 17.06.25 11:31, Dean Rasheed wrote:
On Tue, 17 Jun 2025 at 06:27, Peter Eisentraut <pe...@eisentraut.org> wrote:
Jian is right, there are no OVERRIDING clauses for generated columns,
and the shown code also doesn't check for them, so the comment doesn't
match the code. I suspect that commit 17958972fe3b slightly confused
generated and identity columns in this case. I suggest to remove the
half-sentence "regardless of ...".
Hmm, to me that comment seems quite clear in the context of the
surrounding code. It's making the point that OVERRIDING clauses have
no effect on what can be inserted into generated columns, unlike the
identity column handling code that immediately precedes it.
Perhaps there's an alternate wording that would make that meaning more
explicit, rather than just removing that part of the comment, which
might leave people wondering why that code block isn't checking
"override", when the preceding code blocks do.
I think Jian's point is that the OVERRIDING clause was not created to
affect generated columns, and so mentioning explicitly that it does not
affect generated columns does not seem meaningful. You might as well
mention any other clause there and say it doesn't affect this particular
code.
However, I see your point that it contrasts with the attidentity code
just above.
Perhaps a way to resolve this would be to rewrite the comment something
like:
/*
* Although inserting into a GENERATED BY DEFAULT identity column
* is allowed, apply the default if OVERRIDING USER VALUE is
* specified.
*/
if (att_tup->attidentity == ATTRIBUTE_IDENTITY_BY_DEFAULT &&
override == OVERRIDING_USER_VALUE)
apply_default = true;
/*
* Can only insert DEFAULT into generated columns. (The OVERRIDING
* clause does not apply to generated columns, so we don't consider
* it here.)
*/
if (att_tup->attgenerated && !apply_default)
{
