Re: unnesting multirange data types

On 6/12/21 5:57 PM, Alexander Korotkov wrote:
> On Sat, Jun 12, 2021 at 2:44 AM Alexander Korotkov <aekorotkov(at)gmail(dot)com> wrote:
>> ()On Sat, Jun 12, 2021 at 2:30 AM Justin Pryzby <pryzby(at)telsasoft(dot)com> wrote:
>>> On Fri, Jun 11, 2021 at 11:37:58PM +0300, Alexander Korotkov wrote:
>>>> On Fri, Jun 11, 2021 at 1:04 AM Justin Pryzby <pryzby(at)telsasoft(dot)com> wrote:
>>>>>
>>>>> +{ oid => '1293', descr => 'expand mutlirange to set of ranges',
>>>>>
>>>>> typo: mutlirange
>>>>
>>>> Fixed, thanks.
>>>>
>>>> The patch with the implementation of both unnest() and cast to array
>>>> is attached. It contains both tests and docs.
>>>
>>> |+ The multirange could be explicitly cast to the array of corresponding
>>> should say: "can be cast to an array of corresponding.."
>>>
>>> |+ * Cast multirange to the array of ranges.
>>> I think should be: *an array of ranges
>>
>> Thank you for catching this.
>>
>>> Per sqlsmith, this is causing consistent crashes.
>>> I took one of its less appalling queries and simplified it to this:
>>>
>>> select
>>> pg_catalog.multirange_to_array(
>>> cast(pg_catalog.int8multirange() as int8multirange)) as c2
>>> from (select 1)x;
>>
>> It seems that multirange_to_array() doesn't handle empty multiranges.
>> I'll post an updated version of the patch tomorrow.
>
> A revised patch is attached. Now empty multiranges are handled
> properly (and it's covered by tests). Typos are fixed as well.

Tested both against my original cases using both SQL + PL/pgSQL. All
worked well. I also tested the empty multirange case as well.

Overall the documentation seems to make sense, I'd suggest:

+ <para>
+ The multirange can be cast to an array of corresponding ranges.
+ </para>

becomes:

+ <para>
+ A multirange can be cast to an array of ranges of the same type.
+ </para>

Again, I'll defer to others on the code, but this seems to solve the use
case I presented. Thanks for the quick turnaround!

Jonathan