ARROW-5480: [Python] Add unit test asserting specifically that pandas.Categorical roundtrips to Parquet format without special options

[wesm]

This only works for string types for the moment. Once ARROW-6277 is addressed we can expand to other types.

[wesm]

@jorisvandenbossche@jreback the pandas test suite will probably need some changes or expansions now that Categorical (for strings at least) can be faithfully roundtripped using to_parquet and pd.read_parquet?

[jorisvandenbossche]

@wesm indeed, created pandas-dev/pandas#27955 to track that.

[jorisvandenbossche]

this comment reads a bit unclear to me (it now reads back categorical?)

[wesm]

Sorry I copy-pasted this from the bug report

[jorisvandenbossche]

@wesm to what extent is this "fully" faithful for corner cases? (if not, might need to mention that as caveats in the pandas docs)

For example for a categorical with values in the "categories" which are not present in the data, is this preserved on reading back? (I suppose we use the categories when creating a DictionaryArray, but are its dictionary's values exactly preserved in the parquet roundtrip?)
Or the exact order of the categories in an ordered categorical (if it is not lexicographically sorted) ?

[wesm]

The category values will be exactly preserved whether or not they occur in the data. I can expand the unit test to exhibit this if it helps

[wesm]

Done

[wesm]

Rebased

[jorisvandenbossche]

in the rest of the file, test functions using pandas are marked as such?

[wesm]

Yes, this is a rebase artifact, fixing

[jorisvandenbossche]

I can expand the unit test to exhibit this if it helps

Updated test looks good, thanks for the clarification!

[wesm]

Travis: https://travis-ci.org/wesm/arrow/builds/573882738

[codecov-io]

Codecov Report

Merging #5110 into master will decrease coverage by 22.6%.
The diff coverage is 100%.

@@ Coverage Diff @@## master #5110 +/- ##
===========================================
- Coverage 87.62% 65.02% -22.61% 
===========================================
Files 1014 495 -519 Lines 145828 67082 -78746 Branches 1437 0 -1437 ===========================================
- Hits 127788 43619 -84169 - Misses 17678 23463 +5785 + Partials 362 0 -362

Impacted Files	Coverage Δ
python/pyarrow/tests/test_parquet.py	96.42% <100%> (+0.02%)	⬆️
cpp/src/arrow/util/memory.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/date_utils.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/util/memory.cc	0% <0%> (-100%)	⬇️
cpp/src/arrow/filesystem/util-internal.cc	0% <0%> (-100%)	⬇️
cpp/src/gandiva/decimal_type_util.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/logical_type.h	0% <0%> (-100%)	⬇️
cpp/src/parquet/hasher.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/basic_decimal_scalar.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/kernels/boolean.cc	0% <0%> (-100%)	⬇️
... and 762 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 36bd667...a161b24. Read the comment docs.

[wesm]

This only works for string types for the moment. Once ARROW-6277 is addressed we can expand to other types.

[wesm]

@jorisvandenbossche@jreback the pandas test suite will probably need some changes or expansions now that Categorical (for strings at least) can be faithfully roundtripped using to_parquet and pd.read_parquet?

[jorisvandenbossche]

@wesm indeed, created pandas-dev/pandas#27955 to track that.

[jorisvandenbossche]

this comment reads a bit unclear to me (it now reads back categorical?)

[wesm]

Sorry I copy-pasted this from the bug report

[jorisvandenbossche]

@wesm to what extent is this "fully" faithful for corner cases? (if not, might need to mention that as caveats in the pandas docs)

For example for a categorical with values in the "categories" which are not present in the data, is this preserved on reading back? (I suppose we use the categories when creating a DictionaryArray, but are its dictionary's values exactly preserved in the parquet roundtrip?)
Or the exact order of the categories in an ordered categorical (if it is not lexicographically sorted) ?

[wesm]

The category values will be exactly preserved whether or not they occur in the data. I can expand the unit test to exhibit this if it helps

[wesm]

Done

[wesm]

Rebased

[jorisvandenbossche]

in the rest of the file, test functions using pandas are marked as such?

[wesm]

Yes, this is a rebase artifact, fixing

[jorisvandenbossche]

I can expand the unit test to exhibit this if it helps

Updated test looks good, thanks for the clarification!

[wesm]

Travis: https://travis-ci.org/wesm/arrow/builds/573882738

[codecov-io]

Codecov Report

Merging #5110 into master will decrease coverage by 22.6%.
The diff coverage is 100%.

@@ Coverage Diff @@## master #5110 +/- ##
===========================================
- Coverage 87.62% 65.02% -22.61% 
===========================================
Files 1014 495 -519 Lines 145828 67082 -78746 Branches 1437 0 -1437 ===========================================
- Hits 127788 43619 -84169 - Misses 17678 23463 +5785 + Partials 362 0 -362

Impacted Files	Coverage Δ
python/pyarrow/tests/test_parquet.py	96.42% <100%> (+0.02%)	⬆️
cpp/src/arrow/util/memory.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/date_utils.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/util/memory.cc	0% <0%> (-100%)	⬇️
cpp/src/arrow/filesystem/util-internal.cc	0% <0%> (-100%)	⬇️
cpp/src/gandiva/decimal_type_util.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/logical_type.h	0% <0%> (-100%)	⬇️
cpp/src/parquet/hasher.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/basic_decimal_scalar.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/kernels/boolean.cc	0% <0%> (-100%)	⬇️
... and 762 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 36bd667...a161b24. Read the comment docs.

[wesm]

This only works for string types for the moment. Once ARROW-6277 is addressed we can expand to other types.

[wesm]

@jorisvandenbossche@jreback the pandas test suite will probably need some changes or expansions now that Categorical (for strings at least) can be faithfully roundtripped using to_parquet and pd.read_parquet?

[jorisvandenbossche]

@wesm indeed, created pandas-dev/pandas#27955 to track that.

[jorisvandenbossche]

this comment reads a bit unclear to me (it now reads back categorical?)

[wesm]

Sorry I copy-pasted this from the bug report

[jorisvandenbossche]

@wesm to what extent is this "fully" faithful for corner cases? (if not, might need to mention that as caveats in the pandas docs)

For example for a categorical with values in the "categories" which are not present in the data, is this preserved on reading back? (I suppose we use the categories when creating a DictionaryArray, but are its dictionary's values exactly preserved in the parquet roundtrip?)
Or the exact order of the categories in an ordered categorical (if it is not lexicographically sorted) ?

[wesm]

The category values will be exactly preserved whether or not they occur in the data. I can expand the unit test to exhibit this if it helps

[wesm]

Done

[wesm]

Rebased

[jorisvandenbossche]

in the rest of the file, test functions using pandas are marked as such?

[wesm]

Yes, this is a rebase artifact, fixing

[jorisvandenbossche]

I can expand the unit test to exhibit this if it helps

Updated test looks good, thanks for the clarification!

[wesm]

Travis: https://travis-ci.org/wesm/arrow/builds/573882738

[codecov-io]

Codecov Report

Merging #5110 into master will decrease coverage by 22.6%.
The diff coverage is 100%.

@@ Coverage Diff @@## master #5110 +/- ##
===========================================
- Coverage 87.62% 65.02% -22.61% 
===========================================
Files 1014 495 -519 Lines 145828 67082 -78746 Branches 1437 0 -1437 ===========================================
- Hits 127788 43619 -84169 - Misses 17678 23463 +5785 + Partials 362 0 -362

Impacted Files	Coverage Δ
python/pyarrow/tests/test_parquet.py	96.42% <100%> (+0.02%)	⬆️
cpp/src/arrow/util/memory.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/date_utils.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/util/memory.cc	0% <0%> (-100%)	⬇️
cpp/src/arrow/filesystem/util-internal.cc	0% <0%> (-100%)	⬇️
cpp/src/gandiva/decimal_type_util.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/logical_type.h	0% <0%> (-100%)	⬇️
cpp/src/parquet/hasher.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/basic_decimal_scalar.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/kernels/boolean.cc	0% <0%> (-100%)	⬇️
... and 762 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 36bd667...a161b24. Read the comment docs.

[wesm]

This only works for string types for the moment. Once ARROW-6277 is addressed we can expand to other types.

[wesm]

@jorisvandenbossche@jreback the pandas test suite will probably need some changes or expansions now that Categorical (for strings at least) can be faithfully roundtripped using to_parquet and pd.read_parquet?

[jorisvandenbossche]

@wesm indeed, created pandas-dev/pandas#27955 to track that.

[jorisvandenbossche]

this comment reads a bit unclear to me (it now reads back categorical?)

[wesm]

Sorry I copy-pasted this from the bug report

[jorisvandenbossche]

@wesm to what extent is this "fully" faithful for corner cases? (if not, might need to mention that as caveats in the pandas docs)

For example for a categorical with values in the "categories" which are not present in the data, is this preserved on reading back? (I suppose we use the categories when creating a DictionaryArray, but are its dictionary's values exactly preserved in the parquet roundtrip?)
Or the exact order of the categories in an ordered categorical (if it is not lexicographically sorted) ?

[wesm]

The category values will be exactly preserved whether or not they occur in the data. I can expand the unit test to exhibit this if it helps

[wesm]

Done

[wesm]

Rebased

[jorisvandenbossche]

in the rest of the file, test functions using pandas are marked as such?

[wesm]

Yes, this is a rebase artifact, fixing

[jorisvandenbossche]

I can expand the unit test to exhibit this if it helps

Updated test looks good, thanks for the clarification!

[wesm]

Travis: https://travis-ci.org/wesm/arrow/builds/573882738

[codecov-io]

Codecov Report

Merging #5110 into master will decrease coverage by 22.6%.
The diff coverage is 100%.

@@ Coverage Diff @@## master #5110 +/- ##
===========================================
- Coverage 87.62% 65.02% -22.61% 
===========================================
Files 1014 495 -519 Lines 145828 67082 -78746 Branches 1437 0 -1437 ===========================================
- Hits 127788 43619 -84169 - Misses 17678 23463 +5785 + Partials 362 0 -362

Impacted Files	Coverage Δ
python/pyarrow/tests/test_parquet.py	96.42% <100%> (+0.02%)	⬆️
cpp/src/arrow/util/memory.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/date_utils.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/util/memory.cc	0% <0%> (-100%)	⬇️
cpp/src/arrow/filesystem/util-internal.cc	0% <0%> (-100%)	⬇️
cpp/src/gandiva/decimal_type_util.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/logical_type.h	0% <0%> (-100%)	⬇️
cpp/src/parquet/hasher.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/basic_decimal_scalar.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/kernels/boolean.cc	0% <0%> (-100%)	⬇️
... and 762 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 36bd667...a161b24. Read the comment docs.

[wesm]

This only works for string types for the moment. Once ARROW-6277 is addressed we can expand to other types.

[wesm]

@jorisvandenbossche@jreback the pandas test suite will probably need some changes or expansions now that Categorical (for strings at least) can be faithfully roundtripped using to_parquet and pd.read_parquet?

[jorisvandenbossche]

@wesm indeed, created pandas-dev/pandas#27955 to track that.

[jorisvandenbossche]

this comment reads a bit unclear to me (it now reads back categorical?)

[wesm]

Sorry I copy-pasted this from the bug report

[jorisvandenbossche]

@wesm to what extent is this "fully" faithful for corner cases? (if not, might need to mention that as caveats in the pandas docs)

For example for a categorical with values in the "categories" which are not present in the data, is this preserved on reading back? (I suppose we use the categories when creating a DictionaryArray, but are its dictionary's values exactly preserved in the parquet roundtrip?)
Or the exact order of the categories in an ordered categorical (if it is not lexicographically sorted) ?

[wesm]

The category values will be exactly preserved whether or not they occur in the data. I can expand the unit test to exhibit this if it helps

[wesm]

Done

[wesm]

Rebased

[jorisvandenbossche]

in the rest of the file, test functions using pandas are marked as such?

[wesm]

Yes, this is a rebase artifact, fixing

[jorisvandenbossche]

I can expand the unit test to exhibit this if it helps

Updated test looks good, thanks for the clarification!

[wesm]

Travis: https://travis-ci.org/wesm/arrow/builds/573882738

[codecov-io]

Codecov Report

Merging #5110 into master will decrease coverage by 22.6%.
The diff coverage is 100%.

@@ Coverage Diff @@## master #5110 +/- ##
===========================================
- Coverage 87.62% 65.02% -22.61% 
===========================================
Files 1014 495 -519 Lines 145828 67082 -78746 Branches 1437 0 -1437 ===========================================
- Hits 127788 43619 -84169 - Misses 17678 23463 +5785 + Partials 362 0 -362

Impacted Files	Coverage Δ
python/pyarrow/tests/test_parquet.py	96.42% <100%> (+0.02%)	⬆️
cpp/src/arrow/util/memory.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/date_utils.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/util/memory.cc	0% <0%> (-100%)	⬇️
cpp/src/arrow/filesystem/util-internal.cc	0% <0%> (-100%)	⬇️
cpp/src/gandiva/decimal_type_util.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/logical_type.h	0% <0%> (-100%)	⬇️
cpp/src/parquet/hasher.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/basic_decimal_scalar.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/kernels/boolean.cc	0% <0%> (-100%)	⬇️
... and 762 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 36bd667...a161b24. Read the comment docs.

[wesm]

This only works for string types for the moment. Once ARROW-6277 is addressed we can expand to other types.

[wesm]

@jorisvandenbossche@jreback the pandas test suite will probably need some changes or expansions now that Categorical (for strings at least) can be faithfully roundtripped using to_parquet and pd.read_parquet?

[jorisvandenbossche]

@wesm indeed, created pandas-dev/pandas#27955 to track that.

[jorisvandenbossche]

this comment reads a bit unclear to me (it now reads back categorical?)

[wesm]

Sorry I copy-pasted this from the bug report

[jorisvandenbossche]

@wesm to what extent is this "fully" faithful for corner cases? (if not, might need to mention that as caveats in the pandas docs)

For example for a categorical with values in the "categories" which are not present in the data, is this preserved on reading back? (I suppose we use the categories when creating a DictionaryArray, but are its dictionary's values exactly preserved in the parquet roundtrip?)
Or the exact order of the categories in an ordered categorical (if it is not lexicographically sorted) ?

[wesm]

The category values will be exactly preserved whether or not they occur in the data. I can expand the unit test to exhibit this if it helps

[wesm]

Done

[wesm]

Rebased

[jorisvandenbossche]

in the rest of the file, test functions using pandas are marked as such?

[wesm]

Yes, this is a rebase artifact, fixing

[jorisvandenbossche]

I can expand the unit test to exhibit this if it helps

Updated test looks good, thanks for the clarification!

[wesm]

Travis: https://travis-ci.org/wesm/arrow/builds/573882738

[codecov-io]

Codecov Report

Merging #5110 into master will decrease coverage by 22.6%.
The diff coverage is 100%.

@@ Coverage Diff @@## master #5110 +/- ##
===========================================
- Coverage 87.62% 65.02% -22.61% 
===========================================
Files 1014 495 -519 Lines 145828 67082 -78746 Branches 1437 0 -1437 ===========================================
- Hits 127788 43619 -84169 - Misses 17678 23463 +5785 + Partials 362 0 -362

Impacted Files	Coverage Δ
python/pyarrow/tests/test_parquet.py	96.42% <100%> (+0.02%)	⬆️
cpp/src/arrow/util/memory.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/date_utils.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/util/memory.cc	0% <0%> (-100%)	⬇️
cpp/src/arrow/filesystem/util-internal.cc	0% <0%> (-100%)	⬇️
cpp/src/gandiva/decimal_type_util.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/logical_type.h	0% <0%> (-100%)	⬇️
cpp/src/parquet/hasher.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/basic_decimal_scalar.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/kernels/boolean.cc	0% <0%> (-100%)	⬇️
... and 762 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 36bd667...a161b24. Read the comment docs.

[wesm]

This only works for string types for the moment. Once ARROW-6277 is addressed we can expand to other types.

[wesm]

@jorisvandenbossche@jreback the pandas test suite will probably need some changes or expansions now that Categorical (for strings at least) can be faithfully roundtripped using to_parquet and pd.read_parquet?

[jorisvandenbossche]

@wesm indeed, created pandas-dev/pandas#27955 to track that.

[jorisvandenbossche]

this comment reads a bit unclear to me (it now reads back categorical?)

[wesm]

Sorry I copy-pasted this from the bug report

[jorisvandenbossche]

@wesm to what extent is this "fully" faithful for corner cases? (if not, might need to mention that as caveats in the pandas docs)

For example for a categorical with values in the "categories" which are not present in the data, is this preserved on reading back? (I suppose we use the categories when creating a DictionaryArray, but are its dictionary's values exactly preserved in the parquet roundtrip?)
Or the exact order of the categories in an ordered categorical (if it is not lexicographically sorted) ?

[wesm]

The category values will be exactly preserved whether or not they occur in the data. I can expand the unit test to exhibit this if it helps

[wesm]

Done

[wesm]

Rebased

[jorisvandenbossche]

in the rest of the file, test functions using pandas are marked as such?

[wesm]

Yes, this is a rebase artifact, fixing

[jorisvandenbossche]

I can expand the unit test to exhibit this if it helps

Updated test looks good, thanks for the clarification!

[wesm]

Travis: https://travis-ci.org/wesm/arrow/builds/573882738

[codecov-io]

Codecov Report

Merging #5110 into master will decrease coverage by 22.6%.
The diff coverage is 100%.

@@ Coverage Diff @@## master #5110 +/- ##
===========================================
- Coverage 87.62% 65.02% -22.61% 
===========================================
Files 1014 495 -519 Lines 145828 67082 -78746 Branches 1437 0 -1437 ===========================================
- Hits 127788 43619 -84169 - Misses 17678 23463 +5785 + Partials 362 0 -362

Impacted Files	Coverage Δ
python/pyarrow/tests/test_parquet.py	96.42% <100%> (+0.02%)	⬆️
cpp/src/arrow/util/memory.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/date_utils.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/util/memory.cc	0% <0%> (-100%)	⬇️
cpp/src/arrow/filesystem/util-internal.cc	0% <0%> (-100%)	⬇️
cpp/src/gandiva/decimal_type_util.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/logical_type.h	0% <0%> (-100%)	⬇️
cpp/src/parquet/hasher.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/basic_decimal_scalar.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/kernels/boolean.cc	0% <0%> (-100%)	⬇️
... and 762 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 36bd667...a161b24. Read the comment docs.

[wesm]

This only works for string types for the moment. Once ARROW-6277 is addressed we can expand to other types.

[wesm]

@jorisvandenbossche@jreback the pandas test suite will probably need some changes or expansions now that Categorical (for strings at least) can be faithfully roundtripped using to_parquet and pd.read_parquet?

[jorisvandenbossche]

@wesm indeed, created pandas-dev/pandas#27955 to track that.

[jorisvandenbossche]

this comment reads a bit unclear to me (it now reads back categorical?)

[wesm]

Sorry I copy-pasted this from the bug report

[jorisvandenbossche]

@wesm to what extent is this "fully" faithful for corner cases? (if not, might need to mention that as caveats in the pandas docs)

For example for a categorical with values in the "categories" which are not present in the data, is this preserved on reading back? (I suppose we use the categories when creating a DictionaryArray, but are its dictionary's values exactly preserved in the parquet roundtrip?)
Or the exact order of the categories in an ordered categorical (if it is not lexicographically sorted) ?

[wesm]

The category values will be exactly preserved whether or not they occur in the data. I can expand the unit test to exhibit this if it helps

[wesm]

Done

[wesm]

Rebased

[jorisvandenbossche]

in the rest of the file, test functions using pandas are marked as such?

[wesm]

Yes, this is a rebase artifact, fixing

[jorisvandenbossche]

I can expand the unit test to exhibit this if it helps

Updated test looks good, thanks for the clarification!

[wesm]

Travis: https://travis-ci.org/wesm/arrow/builds/573882738

[codecov-io]

Codecov Report

Merging #5110 into master will decrease coverage by 22.6%.
The diff coverage is 100%.

@@ Coverage Diff @@## master #5110 +/- ##
===========================================
- Coverage 87.62% 65.02% -22.61% 
===========================================
Files 1014 495 -519 Lines 145828 67082 -78746 Branches 1437 0 -1437 ===========================================
- Hits 127788 43619 -84169 - Misses 17678 23463 +5785 + Partials 362 0 -362

Impacted Files	Coverage Δ
python/pyarrow/tests/test_parquet.py	96.42% <100%> (+0.02%)	⬆️
cpp/src/arrow/util/memory.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/date_utils.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/util/memory.cc	0% <0%> (-100%)	⬇️
cpp/src/arrow/filesystem/util-internal.cc	0% <0%> (-100%)	⬇️
cpp/src/gandiva/decimal_type_util.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/logical_type.h	0% <0%> (-100%)	⬇️
cpp/src/parquet/hasher.h	0% <0%> (-100%)	⬇️
cpp/src/gandiva/basic_decimal_scalar.h	0% <0%> (-100%)	⬇️
cpp/src/arrow/compute/kernels/boolean.cc	0% <0%> (-100%)	⬇️
... and 762 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 36bd667...a161b24. Read the comment docs.