Peter Geoghegan wrote:
> In the case of ignore_system_indexes=on, I see the same 17 entries, in
> addition to these 3 (20 total):
>
> drop cascades to table collate_test23 column f1
> drop cascades to table collate_test4 column b
> drop cascades to table collate_test5 column b
>
> Perhaps this means something to you. I find it suspicious that all 3
> possibly-missing entries are "column" entries.
I bet this is related to how are these objects reached while walking the
dependency graph -- i.e. they are reached first as columns and reported
explicitly in the second case, but in the first case the tables are
reached first so the columns are not considered individually. So it'd
just be because of pg_depend scan order.
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services