Remove nvstrdesc_s from cuio

[kaatish]

Fixes #5682.

• Structure nvstrdesc_s was replaced with thrust::pair<const char*, size_type>;.
• nvstrdesc_s related logical functions such as nvstr_is_lesser, nvstr_is_greater etc. were removed.
• Include directives for headers included by source files residing in the same directory were made relative as per the developer guide.
• make_column function related to column_buffer was moved from a header file to an implementation file.

[vuule]

Can you please add a high-level list of changes to the description, and list the issue(s) this PR addresses?

[vuule]

Looks great!
A couple of small asks, just to be as consistent as possible with the rest of libcudf.

[vuule]

@kkraus14 this PR has been incorrectly marked to require Java and Ops reviews. Was there a way to clear this?

[vuule]

👍

[devavret]

🥳

[raydouglass]

Removing ops-codeowners from the required reviews since it doesn't seem there are any file changes that we're responsible for. Feel free to add us back if necessary.

[nvdbaranec]

Looks good to me. I think Devavret's comment on the include style might need to be addressed though.

[jrhemstad]

Why not use string_view instead of thrust::pair<const char*, size_type>?

[vuule]

Why not use string_view instead of thrust::pair<const char*, size_type>?

string_view is larger than the pair because of the size related data members. We really want to avoid memory use increases in these APIs.

[nvdbaranec]

Why not use string_view instead of thrust::pair<const char*, size_type>?

string_view is larger than the pair because of the size related data members. We really want to avoid memory use increases in these APIs.

I don't think that's the case. The whole struct is going to have to be aligned to 8 bytes because of the pointer anyway.

[devavret]

Why not use string_view instead of thrust::pair<const char*, size_type>?

string_view is larger than the pair because of the size related data members. We really want to avoid memory use increases in these APIs.

I don't think that's the case. The whole struct is going to have to be aligned to 8 bytes because of the pointer anyway.

string_view contains 4 members.

cudf/cpp/include/cudf/strings/string_view.hpp

Lines 337 to 340 in ab7fe05

	const char* _data{}; ///< Pointer to device memory contain char array for this string
	size_type _bytes{}; ///< Number of bytes in _data for this string
	mutable size_type _length{}; ///< Number of characters in this string (computed)
	mutable int8_t _char_width{}; ///< Number of bytes per character if uniform width (computed)

2 of these are calculated using the other 2. This is what @vuule is referring to as size related members.

BTW, @jrhemstad, the memory requirement is in column_buffer which is going to allocate a vector of these <data, size> pairs.

[codecov]

Codecov Report

Merging #7841 (efdec2b) into branch-0.20 (51336df) will decrease coverage by 0.37%.
The diff coverage is 87.64%.

❗ Current head efdec2b differs from pull request most recent head 7e5efd7. Consider uploading reports for the commit 7e5efd7 to get more accurate results

@@               Coverage Diff               @@
##           branch-0.20    #7841      +/-   ##
===============================================
- Coverage        82.88%   82.51%   -0.38%     
===============================================
  Files              103      103              
  Lines            17668    17296     -372     
===============================================
- Hits             14645    14272     -373     
- Misses            3023     3024       +1     

Impacted Files	Coverage Δ
python/cudf/cudf/core/column/__init__.py	100.00% <ø> (ø)
python/cudf/cudf/utils/cudautils.py	55.04% <25.00%> (-2.72%)	⬇️
python/cudf/cudf/core/column/numerical.py	94.41% <70.00%> (-0.02%)	⬇️
python/cudf/cudf/core/column/column.py	88.48% <71.42%> (-0.17%)	⬇️
python/cudf/cudf/core/dataframe.py	90.75% <83.33%> (-0.11%)	⬇️
python/cudf/cudf/utils/utils.py	89.47% <91.66%> (-0.04%)	⬇️
python/cudf/cudf/core/column/datetime.py	89.40% <100.00%> (-0.52%)	⬇️
python/cudf/cudf/core/column/timedelta.py	88.33% <100.00%> (-0.34%)	⬇️
python/cudf/cudf/core/groupby/groupby.py	92.12% <100.00%> (+0.67%)	⬆️
python/cudf/cudf/core/index.py	92.62% <100.00%> (-0.46%)	⬇️
... and 53 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 98711da...7e5efd7. Read the comment docs.

[kkraus14]

CMake approval

[devavret]

@gpucibot merge