[
https://issues.apache.org/jira/browse/ARROW-14122?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17420212#comment-17420212
]
Jorge Leitão commented on ARROW-14122:
--------------------------------------
It seems that the question is whether arrow intervals have a [total
order|https://en.wikipedia.org/wiki/Total_order] or [partial
order|https://en.wikipedia.org/wiki/Partially_ordered_set#Non-strict_partial_order].
IMO intervals should not have a total order, only partial order. IMO total
order is only defined for "duration", that has a physical meaning and an
non-overlapping arrow of time.
To get total order, I would suggest people to cast the interval to a duration
(e.g. following postgres' definition of 30 days, 24h, 12 months) and ordering
that. Adopting an assumption over duration of months and days results in loss
of generality, while an extra cast does not and it is more explicit where the
assumption is being made.
Postgres has no concept of duration and thus it makes sense for them to have
both concepts (physical and calendar time differences) mixed together in
intervals.
> [C++] interval comparison kernels
> ---------------------------------
>
> Key: ARROW-14122
> URL: https://issues.apache.org/jira/browse/ARROW-14122
> Project: Apache Arrow
> Issue Type: Sub-task
> Reporter: Phillip Cloud
> Priority: Major
> Labels: kernel
>
> Subtask for tracking interval comparison kernels
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
