=> Should it be called Query_for_list_of_partitioned_tables ? Or should c.relkind match indices, too ?
On Sat, Apr 06, 2019 at 01:36:23AM -0300, Alvaro Herrera wrote: > Maybe the only behavior change I'd do to the submitted patch is to have > \dP show both tables and indexes, while \dPt shows only tables and \dPi > shows only indexes. Maybe have \dPti show both tables and indexes? ( > identical to \dP) That would be consistent with \d itself.
I think there's an issue with showing indices. You said that \dP should be same as \dPti, no? Right now, indices are not shown in \dP, unless a pattern is given. I see you add that behavior in the regression tests; is that really what's intended ? Also, right now adding a pattern affects how sizes are computed, I don't see why that's desirable or, if so, how to resolve that inconsistency, or how to document it.
That depends. If there are not pattern, then \dP show only tables, but with total relation size (so size of indexes are nested). It is different than \dPti, but I think so it is useful - when you don't specify object type, then usually you would to see a tables, but with total size.
I don't see a benefit from \dP == \dPti. When there are a pattern (that can choose some index, then, indexes are displayed and \dP == \dPti.
I think so Alvaro's version is correct, and I prefer it.