út 9. 7. 2019 v 20:25 odesílatel Jeff Davis <pgsql@j-davis.com> napsal:
On Tue, 2019-07-09 at 07:08 +0200, Pavel Stehule wrote: > > I am not against a multirange type, but I miss a explanation why you > introduce new kind of types and don't use just array of ranges. > > Introduction of new kind of types is not like introduction new type.
The biggest benefit, in my opinion, is that it means you can define functions/operators that take an "anyrange" and return an "anymultirange". That way you don't have to define different functions for int4 ranges, date ranges, etc.
I am not sure how strong is this argument.
I think so introduction of anyrangearray polymorphic type and enhancing some type deduction can do same work.
It starts to get even more complex when you want to add opclasses, etc.
Ranges and arrays are effectively generic types that need a type parameter to become a concrete type. Ideally, we'd have first-class support for generic types, but I think that's a different topic ;-)
I afraid so with generic multiragetype there lot of array infrastructure will be duplicated