gabotechs commented on code in PR #26593:
URL: https://github.com/apache/datafusion/pull/26593#discussion_r4489971969
##########
datafusion/physical-expr/src/partitioning.rs:
##########
@@ -217,6 +237,10 @@ pub struct RangePartitioning {
samples: Arc<[SplitPoint]>,
/// Effective boundaries for the current partition count.
split_points: Arc<[SplitPoint]>,
+ /// Inclusive start of the key space, in `ordering` order.
+ start_bound: Option<SplitPoint>,
+ /// Exclusive end of the key space, in `ordering` order.
+ end_bound: Option<SplitPoint>,
Review Comment:
> But as you say, there is no +inf for strings or bytes.
But isn't this still a challenge with the current spec, it's just that now
it's implicit.
🤔 In the case of an empty dataset, wouldn't it be convenient to represent
that somehow rather than returning a (-inf, +inf)? For example, if you know at
planning time there's 0 rows, just replace that node with an `EmptyExec`, and
make Range partitioning over an empty dataset irrepresentable.
In any case, my suggestion here is mainly about cosmetics, I don't think
it's really filling any gap not addressed already accounted for in this PR
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
