Peter Toth comments

Results 118 comments of


                                            Peter Toth

[SPARK-38717][SQL] Handle Hive's bucket spec case preserving behaviour

@HyukjinKwon, @dongjoon-hyun this is an old, small bugfix PR, but it fell into the cracks...

[SPARK-38717][SQL] Handle Hive's bucket spec case preserving behaviour

Thanks for the review! Sure, I will open a 3.3 backport PR today...

[SPARK-40086][SQL] Improve AliasAwareOutputPartitioning to take all aliases into account

cc @cloud-fan

[SPARK-40086][SQL] Improve AliasAwareOutputPartitioning to take all aliases into account

@cloud-fan, @imback82, can you please help to review this PR?

[SPARK-40259][SQL] Support Parquet DSv2 in subquery plan merge

cc @cloud-fan, @sigmod, @singhpk234

[SPARK-45599][CORE] Use object equality in OpenHashSet

Yes, I was watching this ticket because it is a correctness issue. The fix looks good to me. cc @cloud-fan and @HeartSaVioR as it would be great to include the...

[SPARK-45599][CORE] Use object equality in OpenHashSet

> I can hold cutting the tag of RC1 for this. Thanks @HeartSaVioR. BTW I don't think this should be a blocker of 3.5.1 as this is not a regression,...

[SPARK-45599][CORE] Use object equality in OpenHashSet

If we go this direction and change `OpenHashSet` do we still need `SQLOpenHashSet`?

[SPARK-45599][CORE] Use object equality in OpenHashSet

> `SQLOpenHashSet` also handles null differently. Not sure if `OpenHashSet` already covers it. Yes, you are right. Probably the `NaN` special handling can be remvoved though.

Stop copying `Expr`s and LogicalPlans so much during Common Subexpression Elimination

I'm happy to take this issue, but I just noticed that there is a draft PR already: https://github.com/apache/arrow-datafusion/pull/10067. @alamb would you like to finish that PR?