Fix offset_end iterator for lists_column_view, which was not correctl… #7551

ttnghia · 2021-03-10T15:36:11Z

Fix the offset_end iterator in lists_column_view. Since the offset column size is one element larger than the number of column rows, the offset_end should be computed as offset_begin() + size() + 1. This can also be done by offset_begin() + offsets().size().

This PR blocks #7528, thus it must be merged before that PR.

…y computed

cpp/include/cudf/lists/lists_column_view.hpp

Co-authored-by: Jake Hemstad <[email protected]>

codecov · 2021-03-10T19:03:47Z

Codecov Report

Merging #7551 (d95af41) into branch-0.19 (2e4b5a6) will increase coverage by 0.44%.
The diff coverage is n/a.

@@               Coverage Diff               @@
##           branch-0.19    #7551      +/-   ##
===============================================
+ Coverage        81.90%   82.35%   +0.44%     
===============================================
  Files              101      101              
  Lines            16900    17283     +383     
===============================================
+ Hits             13842    14233     +391     
+ Misses            3058     3050       -8

Impacted Files	Coverage Δ
python/cudf/cudf/core/column/decimal.py	`93.33% <0.00%> (-2.50%)`	⬇️
python/cudf/cudf/core/column/numerical.py	`94.85% <0.00%> (-0.17%)`	⬇️
python/cudf/cudf/io/feather.py	`100.00% <0.00%> (ø)`
python/cudf/cudf/comm/serialize.py	`0.00% <0.00%> (ø)`
python/cudf/cudf/_fuzz_testing/io.py	`0.00% <0.00%> (ø)`
python/cudf/cudf/core/column/struct.py	`100.00% <0.00%> (ø)`
python/dask_cudf/dask_cudf/_version.py	`0.00% <0.00%> (ø)`
python/dask_cudf/dask_cudf/io/tests/test_csv.py	`100.00% <0.00%> (ø)`
python/dask_cudf/dask_cudf/io/tests/test_orc.py	`100.00% <0.00%> (ø)`
python/dask_cudf/dask_cudf/io/tests/test_json.py	`100.00% <0.00%> (ø)`
... and 39 more

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 2e4b5a6...d95af41. Read the comment docs.

ttnghia · 2021-03-10T19:59:22Z

@gpucibot merge.

kkraus14 · 2021-03-10T20:01:23Z

@gpucibot merge.

We require two cpp approvals before merging.

kkraus14 · 2021-03-10T20:01:55Z

cc @trxcllnt

ttnghia · 2021-03-10T20:02:38Z

@gpucibot merge.

We require two cpp approvals before merging.

Hah, github tells me "1 approving review" 😄

ttnghia · 2021-03-10T21:34:54Z

@gpucibot merge.

kkraus14 · 2021-03-10T21:35:20Z

Why is this a breaking change? Isn't this a bug fix without API/ABI change?

ttnghia · 2021-03-10T21:36:29Z

Why is this a breaking change? Isn't this just an internal bug fix?

Sorry, it's breaking in my other PR. Maybe I miss understood the breaking meaning.

Hah, without the "breaking" label, CI check fails.

kkraus14 · 2021-03-10T21:37:20Z

Sorry, it's breaking in my other PR. Maybe I miss understood the breaking meaning.

Got it. breaking and non-breaking are referring to downstream consumers of libcudf related to API/ABI/behavior changes that would break their consumption of libcudf.

ttnghia · 2021-03-10T21:38:58Z

@gpucibot merge!!!!!!!!!

Apparently the bot never listens to me.

kkraus14 · 2021-03-10T21:39:39Z

Why is this a breaking change? This is a bug fix for incorrect results with no API or ABI breaking changes?

ttnghia · 2021-03-10T21:40:22Z

Why is this a breaking change? This is a bug fix for incorrect results with no API or ABI breaking changes?

It should not be any breaking change here. But if I remove the breaking label, CI test fails.

ttnghia · 2021-03-10T21:41:00Z

Here we go. Adding non-breaking is also a fix. I didn't know that solution.

kkraus14 · 2021-03-10T21:41:17Z

There's a label checker that checks that either the breaking or non-breaking label is set to indicate the behavior. Please ask if there's uncertainty instead of pushing incorrect information.

kkraus14 · 2021-03-10T21:41:22Z

@gpucibot merge

ttnghia · 2021-03-10T21:41:54Z

There's a label checker that checks that either the breaking or non-breaking label is set to indicate the behavior. Please ask if there's uncertainty instead of pushing incorrect information.

Thanks very much, Keith.

@ttnghia

This is another fix for `offsets_end()` iterator in lists_column_view. The last fix (#7551) was still not correct---that iterator should not be computed using the size of the `offsets()` child column, which is also the offsets of the original (non-sliced) column. Instead, it should be computed using the `size()` of the current column. Interestingly, my previous fix passed all the unit tests, since thrust does not throw anything (like access violation) when the input range is larger than the output range. Authors: - Nghia Truong (@ttnghia) Approvers: - Jake Hemstad (@jrhemstad) - David (@davidwendt) URL: #7575

@ttnghia

rapidsai#7551) Fix the `offset_end` iterator in `lists_column_view`. Since the offset column size is one element larger than the number of column rows, the `offset_end` should be computed as `offset_begin() + size() + 1`. This can also be done by `offset_begin() + offsets().size()`. This PR blocks rapidsai#7528, thus it must be merged before that PR. Authors: - Nghia Truong (@ttnghia) Approvers: - Jake Hemstad (@jrhemstad) - Mike Wilson (@hyperbolic2346) - Vukasin Milovanovic (@vuule) URL: rapidsai#7551

@ttnghia

…#7575) This is another fix for `offsets_end()` iterator in lists_column_view. The last fix (rapidsai#7551) was still not correct---that iterator should not be computed using the size of the `offsets()` child column, which is also the offsets of the original (non-sliced) column. Instead, it should be computed using the `size()` of the current column. Interestingly, my previous fix passed all the unit tests, since thrust does not throw anything (like access violation) when the input range is larger than the output range. Authors: - Nghia Truong (@ttnghia) Approvers: - Jake Hemstad (@jrhemstad) - David (@davidwendt) URL: rapidsai#7575

Fix offset_end iterator for lists_column_view, which was not correctl…

a68ca21

…y computed

ttnghia requested a review from a team as a code owner March 10, 2021 15:36

ttnghia requested review from trxcllnt and jrhemstad March 10, 2021 15:36

ttnghia added libcudf Affects libcudf (C++/CUDA) code. 3 - Ready for Review Ready for review by team 5 - Ready to Merge Testing and reviews complete, ready to merge bug Something isn't working breaking Breaking change labels Mar 10, 2021

jrhemstad reviewed Mar 10, 2021

View reviewed changes

cpp/include/cudf/lists/lists_column_view.hpp Outdated Show resolved Hide resolved

Update cpp/include/cudf/lists/lists_column_view.hpp

d95af41

Co-authored-by: Jake Hemstad <[email protected]>

jrhemstad approved these changes Mar 10, 2021

View reviewed changes

hyperbolic2346 approved these changes Mar 10, 2021

View reviewed changes

ttnghia removed the request for review from trxcllnt March 10, 2021 21:32

ttnghia added the 0 - Blocked Cannot progress due to external reasons label Mar 10, 2021

kkraus14 removed 0 - Blocked Cannot progress due to external reasons 3 - Ready for Review Ready for review by team labels Mar 10, 2021

vuule approved these changes Mar 10, 2021

View reviewed changes

kkraus14 added non-breaking Non-breaking change and removed breaking Breaking change labels Mar 10, 2021

ttnghia removed the non-breaking Non-breaking change label Mar 10, 2021

ttnghia added the breaking Breaking change label Mar 10, 2021

kkraus14 added non-breaking Non-breaking change and removed breaking Breaking change labels Mar 10, 2021

rapids-bot bot merged commit 2d055c3 into rapidsai:branch-0.19 Mar 10, 2021

ttnghia deleted the fix_list_last_offset branch March 10, 2021 22:52

ttnghia mentioned this pull request Mar 11, 2021

Another fix for offsets_end() iterator in lists_column_view #7575

Merged

ttnghia self-assigned this Apr 25, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix offset_end iterator for lists_column_view, which was not correctl… #7551

Fix offset_end iterator for lists_column_view, which was not correctl… #7551

ttnghia commented Mar 10, 2021 •

edited

Loading

codecov bot commented Mar 10, 2021 •

edited

Loading

ttnghia commented Mar 10, 2021

kkraus14 commented Mar 10, 2021 •

edited

Loading

kkraus14 commented Mar 10, 2021

ttnghia commented Mar 10, 2021 •

edited

Loading

ttnghia commented Mar 10, 2021

kkraus14 commented Mar 10, 2021 •

edited

Loading

ttnghia commented Mar 10, 2021 •

edited

Loading

kkraus14 commented Mar 10, 2021

ttnghia commented Mar 10, 2021

kkraus14 commented Mar 10, 2021

ttnghia commented Mar 10, 2021

ttnghia commented Mar 10, 2021

kkraus14 commented Mar 10, 2021

kkraus14 commented Mar 10, 2021

ttnghia commented Mar 10, 2021

Fix offset_end iterator for lists_column_view, which was not correctl… #7551

Fix offset_end iterator for lists_column_view, which was not correctl… #7551

Conversation

ttnghia commented Mar 10, 2021 • edited Loading

codecov bot commented Mar 10, 2021 • edited Loading

Codecov Report

ttnghia commented Mar 10, 2021

kkraus14 commented Mar 10, 2021 • edited Loading

kkraus14 commented Mar 10, 2021

ttnghia commented Mar 10, 2021 • edited Loading

ttnghia commented Mar 10, 2021

kkraus14 commented Mar 10, 2021 • edited Loading

ttnghia commented Mar 10, 2021 • edited Loading

kkraus14 commented Mar 10, 2021

ttnghia commented Mar 10, 2021

kkraus14 commented Mar 10, 2021

ttnghia commented Mar 10, 2021

ttnghia commented Mar 10, 2021

kkraus14 commented Mar 10, 2021

kkraus14 commented Mar 10, 2021

ttnghia commented Mar 10, 2021

ttnghia commented Mar 10, 2021 •

edited

Loading

codecov bot commented Mar 10, 2021 •

edited

Loading

kkraus14 commented Mar 10, 2021 •

edited

Loading

ttnghia commented Mar 10, 2021 •

edited

Loading

kkraus14 commented Mar 10, 2021 •

edited

Loading

ttnghia commented Mar 10, 2021 •

edited

Loading