Prepare arrow 52.0.0

[tustvold]

Which issue does this PR close?

Closes #5688

Rationale for this change

What changes are included in this PR?

Are there any user-facing changes?

[alamb]

Thank you @tustvold

[alamb]

A substantial and distinguished release

[tustvold]

I was intending to hold this until I had updated DataFusion...

[tustvold]

I've gone through apache/datafusion#10765 and the remaining test failures are mostly either formatting or memory usage changes. Thank you @waynexia

The one exception to this are the symmetric hash join, but given the amount of interval-specific logic in DF for these I am inclined to think the issue likely does not lie in arrow-rs.

As such I am going to proceed with the release, worst case we have to cut a new RC or do a fast follow patch release which would not be the end of the world.

[waynexia]

Thanks, I haven't meet the symmetric hash join under the default feature gate, hope I can solve the remaining failures tomorrow and the release can go smoothly on both sides. Raphael Taylor-Davies ***@***.***>

…

I've gone through apache/datafusion#10765 <apache/datafusion#10765> and the remaining test failures are mostly either formatting or memory usage changes. Thank you @waynexia <https://github.com/waynexia> The one exception to this are the symmetric hash join, but given the amount of interval-specific logic in DF for these I am inclined to think the issue likely does not lie in arrow-rs. As such I am going to proceed with the release, worst case we have to cut a new RC or do a fast follow patch release which would not be the end of the world. — Reply to this email directly, view it on GitHub <#5836 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADVK7M33MO6LQ2WBISMMCGDZFSGSZAVCNFSM6AAAAABIWHUPYCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNBVGU3DSMBXGA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[waynexia]

The one exception to this are the symmetric hash join, but given the amount of interval-specific logic in DF for these I am inclined to think the issue likely does not lie in arrow-rs.

The fix commit is apache/datafusion@fc9a86b. I'm not sure if IntervalDayTime::ONE should be "1 day 1 milli" or "0 day 1 milli" 🤔 But this should not be a big problem as we can choose not to use that const if its value is not expected. Only concern about it may bring some misunderstanding.

[tustvold]

I'm not sure if IntervalDayTime::ONE should be "1 day 1 milli"

ONE is defined as the multiplicative identity, which in this case is 1 day 1 milli. I agree there is definitely room for confusion when it comes to arithmetic on these types