On Tue, Feb 2, 2021 at 6:49 PM Peter Eisentraut <peter.eisentr...@2ndquadrant.com> wrote: > > I do wonder, however, under what circumstances code would be put into a > situation where it would add the exact same dependency again, and also > under what circumstances it would add a dependency between the same > objects but a different deptype, and how that would work during > recursive deletion. Now that we have the refobjversion column, the > presence of duplicate dependencies might be even more dubious. I think > that could be worth analyzing.
There's at least dependencies from indexes to pg_class (underlying table) and pg_collation that can be entirely duplicated (if the same column and/or collation is used in both key and predicate for instance), and this was the case before refobjversion was introduced.