Sort binary

[waynexia]

Which issue does this PR close?

Closes #568.

What changes are included in this PR?

• Extend take kernel to binary array family
• Implement sort for binary array family
• Unit test

[codecov-commenter]

Codecov Report

Merging #569 (5877bad) into master (1d6c374) will increase coverage by 0.07%.
The diff coverage is 98.55%.

@@            Coverage Diff             @@
##           master     #569      +/-   ##
==========================================
+ Coverage   82.41%   82.49%   +0.07%     
==========================================
  Files         167      167              
  Lines       46257    46465     +208     
==========================================
+ Hits        38125    38331     +206     
- Misses       8132     8134       +2     

Impacted Files	Coverage Δ
arrow/src/compute/kernels/sort.rs	94.69% <98.35%> (+0.54%)	⬆️
arrow/src/array/cast.rs	85.71% <100.00%> (+2.38%)	⬆️
arrow/src/compute/kernels/take.rs	94.91% <100.00%> (+0.18%)	⬆️
parquet/src/encodings/encoding.rs	95.04% <0.00%> (+0.19%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 1d6c374...5877bad. Read the comment docs.

[alamb]

Thank you @waynexia

I am not super familiar with the internals of list generic binary array, but I reviewed the tests you have added (thank you!) and they seem to demonstrate the correct behavior.

@nevi-me or @jorgecarleitao might you have some time to review this PR?

[alamb]

Unless I hear anything else I plan to merge this later today or tomorrow

[alamb]

Thanks again @waynexia 🚀

[waynexia]

Appreciate it @alamb!

[alamb]

BTW @waynexia I expect this change to be available in arrow 5.1.0 (scheduled for release at the end of this week or early next week)

[waynexia]

Sure! I'd like to help.

It seems that changes in #552 make the cherry-pick PR's CI failed. I think #552 should be cherry-picked first.

However, #552 changes the default behavior (from stable sort to unstable version). So if ship #552 in a minor release is not expected, I can submit a patch to make #621 pass CI.

[waynexia]

Ahh, I noticed #573 which already gives a try.

May I ask which way should I add changes to #621? Add commits to branch cherry_pick_2113e2b1 (if I can) or file a new PR.

[alamb]

Thank you very much @waynexia !

May I ask which way should I add changes to #621? Add commits to branch cherry_pick_2113e2b1 (if I can) or file a new PR.

I think a new PR would be best (feel free to base it off of cherry_pick_2113e2b1) so that it is clear what is happening

[waynexia]

Got it, I opened #624 😃