On Tue, Aug 4, 2026 at 6:26 PM Chao Li <[email protected]> wrote: > > > > > On Aug 4, 2026, at 13:11, Peter Smith <[email protected]> wrote: > > > > Here's a patch to add another tab-completion rule for DROP PROPERTY GRAPH > > [1] > > > > The logic is the same as DROP MATERIALIZED VIEW [2] which has a very > > similar syntax. > > > > PSA v1. > > > > ====== > > [1] https://www.postgresql.org/docs/19/sql-drop-property-graph.html > > [2] https://www.postgresql.org/docs/19/sql-dropmaterializedview.html > > > > Kind Regards, > > Peter Smith. > > Fujitsu Australia > > <v1-0001-tab-complete-for-DROP-PROPERTY-GRAPH.patch> > > I think you also need to handle “IF EXISTS”. With v1, MatchAny considers “IF” > as a relation name, thus it suggests "CASCADE RESTRICT”, and when “IF > EXISTS” is given, it doesn’t suggest anything: > ``` > evantest=# drop property graph if > CASCADE RESTRICT > evantest=# drop property graph if exists a >
Yeah, this patch only added CASECADE|RESTRICT in a way that's compatible with many of the other DROPS... (As mentioned in the original post, I copied the same rules as seen in DROP MATERIALIZED VIEW) So, what you reported is definitely a problem, but IMO it's a separate problem beyond the scope for this little patch. There's a lot of existing code which won't behave correctly for your IF EXISTS... e.g. Try any of these: - DROP MATERIALIZED VIEW IF <tab> - DROP INDEX IF <tab> - DROP ACCESS METHOD IF <tab> - DROP EVENT TRIGGER IF <tab> etc... ====== Kind Regards, Peter Smith. Fujitsu Australia
