Remove is_relationally_comparable for table device views

[davidwendt]

Closes #10080

Removes is_relationally_comparable function for table_device_view and mutable_table_device_view. The existing is_relationally_comparable can still be used with non-device table_view and mutable_table_view.

Callers are updated as appropriate. The incorrect kernel call has been removed.

[jrhemstad]

I am very confused about why this code needs a stream or run on the device at all.

[codecov]

Codecov Report

Merging #10342 (66a3124) into branch-22.04 (289a6a1) will increase coverage by 0.01%.
The diff coverage is n/a.

@@               Coverage Diff                @@
##           branch-22.04   #10342      +/-   ##
================================================
+ Coverage         86.13%   86.15%   +0.01%     
================================================
  Files               139      139              
  Lines             22465    22460       -5     
================================================
  Hits              19351    19351              
+ Misses             3114     3109       -5     

Impacted Files	Coverage Δ
python/cudf/cudf/core/column/datetime.py	89.07% <0.00%> (-0.05%)	⬇️
python/cudf/cudf/core/single_column_frame.py	97.01% <0.00%> (-0.03%)	⬇️
python/cudf/cudf/core/column/column.py	88.99% <0.00%> (-0.02%)	⬇️
python/cudf/cudf/core/frame.py	91.72% <0.00%> (-0.01%)	⬇️
python/cudf/cudf/core/column/string.py	88.39% <0.00%> (+0.10%)	⬆️
python/cudf/cudf/core/tools/datetimes.py	84.49% <0.00%> (+0.30%)	⬆️
python/cudf/cudf/core/groupby/groupby.py	91.92% <0.00%> (+0.43%)	⬆️
python/cudf/cudf/core/column/lists.py	90.56% <0.00%> (+0.47%)	⬆️

---

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 7ff1956...66a3124. Read the comment docs.

[vyasr]

@davidwendt just FYI I have avoided reviewing this PR so far because I'm not clear on whether @jrhemstad's question is resolved or whether to expect a real resolution within the scope of this PR, and it seems like a decision there could significantly alter this changeset. Please let me know if you think I should just go ahead and review.

[davidwendt]

@davidwendt just FYI I have avoided reviewing this PR so far because I'm not clear on whether @jrhemstad's question is resolved or whether to expect a real resolution within the scope of this PR, and it seems like a decision there could significantly alter this changeset. Please let me know if you think I should just go ahead and review.

Thanks @vyasr
I think the right approach here is to remove the troublesome is_relationally_comparable_functor for table device-views. This will remove the incorrect kernel/thrust call and solve the original issue. I believe @devavret will be wrapping the row_lexicographic_comparator with an owning class in a separate PR (#10164) which could re-enable the comparable checking on non-device table views there instead.

I welcome all and any reviews at this time.

[devavret]

I believe @devavret will be wrapping the row_lexicographic_comparator with an owning class in a separate PR (#10164) which could re-enable the comparable checking on non-device table views there instead.

That is correct. The check will then be on table_view. But #10164 will have the new comparator in namespace experimental and won't immediately replace the existing comparator. So there may be benefit in this PR going through now.

[vyasr]

LGTM.

[ttnghia]

extern should not be used here. As I found https://stackoverflow.com/questions/8130602/using-extern-template-c11):

You should only use extern template to force the compiler to not instantiate a template when you know that it will be instantiated somewhere else. It is used to reduce compile time and object file size.

[davidwendt]

@gpucibot merge