[multistage] Handle Integer.MIN_VALUE in FieldSelectionKeySelector

[ankitsultana]

Saw this in our production where one of the inputs hashes to Integer.MIN_VALUE.

Math.abs returns negation of the input number if the input number is negative. So for Integer.MIN_VALUE, it returns the same value. This is also documented in the docs for Math.abs.

In other places like MurmurPartitionFunction, HashCodePartitionFunction, etc. we seem to be already doing this so looks like it's a known thing.

[codecov-commenter]

Codecov Report

Merging #10596 (e585755) into master (7c1d10a) will decrease coverage by 41.29%.
The diff coverage is 0.00%.

@@              Coverage Diff              @@
##             master   #10596       +/-   ##
=============================================
- Coverage     69.04%   27.75%   -41.29%     
+ Complexity     6496       58     -6438     
=============================================
  Files          2106     2090       -16     
  Lines        113004   112572      -432     
  Branches      17026    16971       -55     
=============================================
- Hits          78021    31243    -46778     
- Misses        29495    78213    +48718     
+ Partials       5488     3116     -2372     

Flag	Coverage Δ
integration1	24.46% <0.00%> (?)
integration2	24.04% <0.00%> (-0.11%)	⬇️
unittests1	?
unittests2	?

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...lanner/partitioning/FieldSelectionKeySelector.java	0.00% <0.00%> (-60.00%)	⬇️

... and 1550 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[yupeng9]

LG. can you check test failure?

[ankitsultana]

LG. can you check test failure?

Thanks. Filed a separate issue for it: #10597

It might be a flaky test. The code change in this is unrelated since FieldSelectionKeySelector is only used for multistage queries during query execution, and the error in the UT was from some pulsar consumer test.