Re: pg_dump insert column GENERATED

[Дмитрий Иванов <firstdismay@xxxxxxxxx>]

Ok, I'll take a look. 

I recreated the is_inside column yesterday, but it didn't make any difference.

Uchet=# select attname, attgenerated from pg_attribute
Uchet-# where attrelid = 'bpd.object'::regclass and attnum > 0;
            attname            | attgenerated
-------------------------------+--------------
 id                            |
 id_class                      |
 id_position                   |
 ........pg.dropped.4........  |
 ........pg.dropped.5........  |
 ........pg.dropped.6........  |
 bquantity                     |
 ........pg.dropped.8........  |
 ........pg.dropped.9........  |
 id_position_root              |
 id_conception                 |
 barcode_unit                  |
 id_unit_conversion_rule       |
 ........pg.dropped.14........ |
 timestamp                     |
 on_freeze                     |
 timestamp_class               |
 name                          |
 id_class_root                 |
 id_group                      |
 id_group_root                 |
 id_object_carrier             |
 desc                          |
 ........pg.dropped.24........ |
 ........pg.dropped.25........ |
 ........pg.dropped.26........ |
 id_class_prop_object_carrier  |
 id_pos_temp_prop              |
 ........pg.dropped.29........ |
 mc                            |
 is_inside                     | s

вс, 21 нояб. 2021 г. в 21:24, Tom Lane <tgl@xxxxxxxxxxxxx>:

=?UTF-8?B?0JTQvNC40YLRgNC40Lkg0JjQstCw0L3QvtCy?= <firstdismay@xxxxxxxxx> writes:
>> So did you do an ALTER TABLE ... GENERATED ALWAYS AS on bpg.object in
>> the Postgres 12 version of the database?

> I think you're right, I added later.

Hmm, that's suggestive, but I tried making a table that way and still did
not see any pg_dump misbehavior.

It might be useful to see the results of this query on the v12 server:

select attname, attgenerated from pg_attribute
where attrelid = 'bpd.object'::regclass and attnum > 0;

We should see attgenerated = 's' for the generated column, but
maybe we don't?

                        regards, tom lane