rpms
/
python-pandas
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Backport proper pyarrow 10 and 11 support

This commit is contained in:
Benjamin A. Beasley 2023-04-19 08:47:23 -04:00
parent 0cccc16aae
commit 25cbd2676d
2 changed files with 373 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

335
pandas-1.5.3-pyarrow-10-11.patch Normal file
View File

@ -0,0 +1,335 @@
From 43d4450e7e7386eb3aebb286b6101889c32ba52c Mon Sep 17 00:00:00 2001
From: Patrick Hoefler <61934744+phofl@users.noreply.github.com>
Date: Sat, 24 Dec 2022 20:49:35 +0100
Subject: [PATCH 1/3] CI: Unpin pyarrow<10 (#50314)
* CI: Unpin pyarrow<10
* Skip test
---
.github/actions/setup-conda/action.yml | 2 +-
.github/workflows/ubuntu.yml | 2 +-
ci/deps/actions-310.yaml | 2 +-
ci/deps/actions-38-downstream_compat.yaml | 2 +-
ci/deps/actions-38.yaml | 2 +-
ci/deps/actions-39.yaml | 2 +-
ci/deps/circle-38-arm64.yaml | 2 +-
environment.yml | 2 +-
pandas/compat/pyarrow.py | 2 ++
pandas/tests/arrays/interval/test_interval.py | 2 +-
pandas/tests/arrays/period/test_arrow_compat.py | 3 +++
requirements-dev.txt | 2 +-
12 files changed, 15 insertions(+), 10 deletions(-)
diff --git a/.github/actions/setup-conda/action.yml b/.github/actions/setup-conda/action.yml
index 7d1e54052f..002d0020c2 100644
--- a/.github/actions/setup-conda/action.yml
+++ b/.github/actions/setup-conda/action.yml
@@ -18,7 +18,7 @@ runs:
- name: Set Arrow version in ${{ inputs.environment-file }} to ${{ inputs.pyarrow-version }}
run: |
grep -q ' - pyarrow' ${{ inputs.environment-file }}
- sed -i"" -e "s/ - pyarrow<10/ - pyarrow=${{ inputs.pyarrow-version }}/" ${{ inputs.environment-file }}
+ sed -i"" -e "s/ - pyarrow/ - pyarrow=${{ inputs.pyarrow-version }}/" ${{ inputs.environment-file }}
cat ${{ inputs.environment-file }}
shell: bash
if: ${{ inputs.pyarrow-version }}
diff --git a/.github/workflows/ubuntu.yml b/.github/workflows/ubuntu.yml
index 4602d12d85..8fd69a4851 100644
--- a/.github/workflows/ubuntu.yml
+++ b/.github/workflows/ubuntu.yml
@@ -31,7 +31,7 @@ jobs:
pattern: ["not single_cpu", "single_cpu"]
# Don't test pyarrow v2/3: Causes timeouts in read_csv engine
# even if tests are skipped/xfailed
- pyarrow_version: ["5", "6", "7"]
+ pyarrow_version: ["5", "6", "7", "8", "9", "10"]
include:
- name: "Downstream Compat"
env_file: actions-38-downstream_compat.yaml
diff --git a/ci/deps/actions-310.yaml b/ci/deps/actions-310.yaml
index deb23d435b..6050a28e11 100644
--- a/ci/deps/actions-310.yaml
+++ b/ci/deps/actions-310.yaml
@@ -39,7 +39,7 @@ dependencies:
- psycopg2
- pymysql
- pytables
- - pyarrow<10
+ - pyarrow
- pyreadstat
- python-snappy
- pyxlsb
diff --git a/ci/deps/actions-38-downstream_compat.yaml b/ci/deps/actions-38-downstream_compat.yaml
index 06ffafeb70..988eacdd0c 100644
--- a/ci/deps/actions-38-downstream_compat.yaml
+++ b/ci/deps/actions-38-downstream_compat.yaml
@@ -38,7 +38,7 @@ dependencies:
- odfpy
- pandas-gbq
- psycopg2
- - pyarrow<10
+ - pyarrow
- pymysql
- pyreadstat
- pytables
diff --git a/ci/deps/actions-38.yaml b/ci/deps/actions-38.yaml
index 222da40ea9..131e2d1882 100644
--- a/ci/deps/actions-38.yaml
+++ b/ci/deps/actions-38.yaml
@@ -37,7 +37,7 @@ dependencies:
- odfpy
- pandas-gbq
- psycopg2
- - pyarrow<10
+ - pyarrow
- pymysql
- pyreadstat
- pytables
diff --git a/ci/deps/actions-39.yaml b/ci/deps/actions-39.yaml
index 1c60e8ad6d..62e7397757 100644
--- a/ci/deps/actions-39.yaml
+++ b/ci/deps/actions-39.yaml
@@ -38,7 +38,7 @@ dependencies:
- pandas-gbq
- psycopg2
- pymysql
- - pyarrow<10
+ - pyarrow
- pyreadstat
- pytables
- python-snappy
diff --git a/ci/deps/circle-38-arm64.yaml b/ci/deps/circle-38-arm64.yaml
index 263521fb74..512c47f0a6 100644
--- a/ci/deps/circle-38-arm64.yaml
+++ b/ci/deps/circle-38-arm64.yaml
@@ -37,7 +37,7 @@ dependencies:
- odfpy
- pandas-gbq
- psycopg2
- - pyarrow<10
+ - pyarrow
- pymysql
# Not provided on ARM
#- pyreadstat
diff --git a/environment.yml b/environment.yml
index 20f839db9a..1620bad9b0 100644
--- a/environment.yml
+++ b/environment.yml
@@ -38,7 +38,7 @@ dependencies:
- odfpy
- pandas-gbq
- psycopg2
- - pyarrow<10
+ - pyarrow
- pymysql
- pyreadstat
- pytables
diff --git a/pandas/compat/pyarrow.py b/pandas/compat/pyarrow.py
index 6965865acb..887ae49c3d 100644
--- a/pandas/compat/pyarrow.py
+++ b/pandas/compat/pyarrow.py
@@ -18,6 +18,7 @@ try:
pa_version_under7p0 = _palv < Version("7.0.0")
pa_version_under8p0 = _palv < Version("8.0.0")
pa_version_under9p0 = _palv < Version("9.0.0")
+ pa_version_under10p0 = _palv < Version("10.0.0")
except ImportError:
pa_version_under1p01 = True
pa_version_under2p0 = True
@@ -28,3 +29,4 @@ except ImportError:
pa_version_under7p0 = True
pa_version_under8p0 = True
pa_version_under9p0 = True
+ pa_version_under10p0 = True
diff --git a/pandas/tests/arrays/interval/test_interval.py b/pandas/tests/arrays/interval/test_interval.py
index 2a6bea3255..0667082784 100644
--- a/pandas/tests/arrays/interval/test_interval.py
+++ b/pandas/tests/arrays/interval/test_interval.py
@@ -287,7 +287,7 @@ def test_arrow_array():
with pytest.raises(TypeError, match="Not supported to convert IntervalArray"):
pa.array(intervals, type="float64")
- with pytest.raises(TypeError, match="different 'subtype'"):
+ with pytest.raises(TypeError, match="different 'subtype'|to convert IntervalArray"):
pa.array(intervals, type=ArrowIntervalType(pa.float64(), "left"))
diff --git a/pandas/tests/arrays/period/test_arrow_compat.py b/pandas/tests/arrays/period/test_arrow_compat.py
index 03fd146572..01a97d07a7 100644
--- a/pandas/tests/arrays/period/test_arrow_compat.py
+++ b/pandas/tests/arrays/period/test_arrow_compat.py
@@ -1,5 +1,7 @@
import pytest
+from pandas.compat.pyarrow import pa_version_under10p0
+
from pandas.core.dtypes.dtypes import PeriodDtype
import pandas as pd
@@ -26,6 +28,7 @@ def test_arrow_extension_type():
assert not hash(p1) == hash(p3)
+@pytest.mark.xfail(not pa_version_under10p0, reason="Wrong behavior with pyarrow 10")
@pytest.mark.parametrize(
"data, freq",
[
diff --git a/requirements-dev.txt b/requirements-dev.txt
index 95291e4ab5..1c7a011e5f 100644
--- a/requirements-dev.txt
+++ b/requirements-dev.txt
@@ -29,7 +29,7 @@ openpyxl
odfpy
pandas-gbq
psycopg2
-pyarrow<10
+pyarrow
pymysql
pyreadstat
tables
--
2.39.2
From 5c2ced8f67fb248d6e5166b5dfdb03909de3123b Mon Sep 17 00:00:00 2001
From: "Benjamin A. Beasley" <code@musicinmybrain.net>
Date: Wed, 19 Apr 2023 11:36:21 -0400
Subject: [PATCH 2/3] Add pandas.compat.pa_version_under11p0
Partial backport of #50998 / 52306d9
---
pandas/compat/__init__.py | 2 ++
pandas/compat/pyarrow.py | 2 ++
2 files changed, 4 insertions(+)
diff --git a/pandas/compat/__init__.py b/pandas/compat/__init__.py
index 80f66c945b..68f2495d0d 100644
--- a/pandas/compat/__init__.py
+++ b/pandas/compat/__init__.py
@@ -29,6 +29,7 @@ from pandas.compat.pyarrow import (
pa_version_under7p0,
pa_version_under8p0,
pa_version_under9p0,
+ pa_version_under11p0,
)
if TYPE_CHECKING:
@@ -163,4 +164,5 @@ __all__ = [
"pa_version_under7p0",
"pa_version_under8p0",
"pa_version_under9p0",
+ "pa_version_under11p0",
]
diff --git a/pandas/compat/pyarrow.py b/pandas/compat/pyarrow.py
index 887ae49c3d..ca51d74828 100644
--- a/pandas/compat/pyarrow.py
+++ b/pandas/compat/pyarrow.py
@@ -19,6 +19,7 @@ try:
pa_version_under8p0 = _palv < Version("8.0.0")
pa_version_under9p0 = _palv < Version("9.0.0")
pa_version_under10p0 = _palv < Version("10.0.0")
+ pa_version_under11p0 = _palv < Version("11.0.0")
except ImportError:
pa_version_under1p01 = True
pa_version_under2p0 = True
@@ -30,3 +31,4 @@ except ImportError:
pa_version_under8p0 = True
pa_version_under9p0 = True
pa_version_under10p0 = True
+ pa_version_under11p0 = True
--
2.39.2
From 8549bbe4e1144e2429612fa17e082ab9c1cba23f Mon Sep 17 00:00:00 2001
From: Luke Manley <lukemanley@gmail.com>
Date: Fri, 10 Feb 2023 13:15:08 -0500
Subject: [PATCH 3/3] CI: unpin pyarrow, fix failing test (#51175)
* unpin pyarrow, fix failing test
* cleanup
* handle NaT/NaN
---
pandas/core/arrays/arrow/array.py | 12 ++++++++++++
pandas/core/tools/timedeltas.py | 4 +++-
pandas/tests/extension/test_arrow.py | 12 +++++++++++-
3 files changed, 26 insertions(+), 2 deletions(-)
diff --git a/pandas/core/arrays/arrow/array.py b/pandas/core/arrays/arrow/array.py
index 4dfd8942c2..7031f4d80d 100644
--- a/pandas/core/arrays/arrow/array.py
+++ b/pandas/core/arrays/arrow/array.py
@@ -259,6 +259,18 @@ class ArrowExtensionArray(OpsMixin, ExtensionArray):
from pandas.core.tools.timedeltas import to_timedelta
scalars = to_timedelta(strings, errors="raise")
+ if pa_type.unit != "ns":
+ # GH51175: test_from_sequence_of_strings_pa_array
+ # attempt to parse as int64 reflecting pyarrow's
+ # duration to string casting behavior
+ mask = isna(scalars)
+ if not isinstance(strings, (pa.Array, pa.ChunkedArray)):
+ strings = pa.array(strings, type=pa.string(), from_pandas=True)
+ strings = pc.if_else(mask, None, strings)
+ try:
+ scalars = strings.cast(pa.int64())
+ except pa.ArrowInvalid:
+ pass
elif pa.types.is_time(pa_type):
from pandas.core.tools.times import to_time
diff --git a/pandas/core/tools/timedeltas.py b/pandas/core/tools/timedeltas.py
index 5026c97c0b..67ba7818d1 100644
--- a/pandas/core/tools/timedeltas.py
+++ b/pandas/core/tools/timedeltas.py
@@ -236,7 +236,9 @@ def _convert_listlike(arg, unit=None, errors="raise", name=None):
# returning arg (errors == "ignore"), and where the input is a
# generator, we return a useful list-like instead of a
# used-up generator
- arg = np.array(list(arg), dtype=object)
+ if not hasattr(arg, "__array__"):
+ arg = list(arg)
+ arg = np.array(arg, dtype=object)
try:
td64arr = sequence_to_td64ns(arg, unit=unit, errors=errors, copy=False)[0]
diff --git a/pandas/tests/extension/test_arrow.py b/pandas/tests/extension/test_arrow.py
index b7ddb1f248..689bd29dfd 100644
--- a/pandas/tests/extension/test_arrow.py
+++ b/pandas/tests/extension/test_arrow.py
@@ -30,6 +30,7 @@ from pandas.compat import (
pa_version_under7p0,
pa_version_under8p0,
pa_version_under9p0,
+ pa_version_under11p0,
)
from pandas.errors import PerformanceWarning
@@ -272,7 +273,7 @@ class TestConstructors(base.BaseConstructorsTests):
reason="Nanosecond time parsing not supported.",
)
)
- elif pa.types.is_duration(pa_dtype):
+ elif pa_version_under11p0 and pa.types.is_duration(pa_dtype):
request.node.add_marker(
pytest.mark.xfail(
raises=pa.ArrowNotImplementedError,
@@ -1707,3 +1708,12 @@ def test_mode(data_for_grouping, dropna, take_idx, exp_idx, request):
result = ser.mode(dropna=dropna)
expected = pd.Series(data_for_grouping.take(exp_idx))
tm.assert_series_equal(result, expected)
+
+@pytest.mark.parametrize("unit", ["ns", "us", "ms", "s"])
+def test_duration_from_strings_with_nat(unit):
+ # GH51175
+ strings = ["1000", "NaT"]
+ pa_type = pa.duration(unit)
+ result = ArrowExtensionArray._from_sequence_of_strings(strings, dtype=pa_type)
+ expected = ArrowExtensionArray(pa.array([1000, None], type=pa_type))
+ tm.assert_extension_array_equal(result, expected)
--
2.39.2

48
python-pandas.spec
View File

@ -94,6 +94,43 @@ Patch: https://github.com/pandas-dev/pandas/commit/e73d4d29203dab20e001
# https://bugzilla.redhat.com/show_bug.cgi?id=2171682
Patch: https://github.com/pandas-dev/pandas/pull/52150.patch
# CI: Unpin pyarrow<10
# https://github.com/pandas-dev/pandas/pull/50314
#
# Merged upstream as 4878dfe551da2fa8e2bc33e774b595f099bfa74e:
# CI: Unpin pyarrow<10 (#50314)
# * CI: Unpin pyarrow<10
# * Skip test
#
# ----
#
# Add pandas.compat.pa_version_under11p0, required for pyarrow 11 support, from
# the following PR/commit, but without the other associated changes:
#
# ENH: support reductions for pyarrow temporal types (#50998)
# https://github.com/pandas-dev/pandas/pull/50998
#
# Merged upstream as 52306d957cb77a3823624679bb9606e244e7faa8:
# ENH: support reductions for pyarrow temporal types
# * unit check
# * lint fixup
#
# ----
#
# CI: unpin pyarrow, fix failing test
# https://github.com/pandas-dev/pandas/pull/51175
#
# Merged upstream as 5f584bd29be7203db64bdf8619991927e29c74bc:
# CI: unpin pyarrow, fix failing test (#51175)
# * unpin pyarrow, fix failing test
# * cleanup
# * handle NaT/NaN
#
# ----
#
# All commits cherry-picked to tag v1.5.3 and combined into a single patch.
Patch: pandas-1.5.3-pyarrow-10-11.patch
%global _description %{expand:
pandas is an open source, BSD-licensed library providing
high-performance, easy-to-use data structures and data
@ -623,16 +660,6 @@ k="${k-}${k+ and }not (TestDataFramePlotsSubplots and test_bar_log_subplots)"
%endif
%if 0%{?fedora} > 37
# The text of an error message has changed in libarrow/pyarrow 10, which is
# harmless but breaks one test. Disable it until a patch is available upstream.
# CI: pyarrow 10 broke our ci
# https://github.com/pandas-dev/pandas/issues/50058
k="${k-}${k+ and }not test_arrow_array"
# Probably also related to upstream pinning pyarrow < 10 for CI:
# E TypeError: Expected unicode, got pyarrow.lib.StringScalar
k="${k-}${k+ and }not (TestConstructors and test_from_sequence_of_strings_pa_array)"
# TODO: Why does this fail? Does it need a slightly older version of dask?
# E AssertionError: Caused unexpected warning(s): [('RuntimeWarning', RuntimeWarning('invalid value encountered in cast'), '/builddir/build/BUILDROOT/python-pandas-1.5.3-1.fc39.x86_64/usr/lib64/python3.11/site-packages/pandas/core/dtypes/cast.py', 1836)]
k="${k-}${k+ and }not test_construct_dask_float_array_int_dtype_match_ndarray"
@ -717,6 +744,7 @@ export PYTHONHASHSEED="$(
%changelog
* Wed Apr 19 2023 Benjamin A. Beasley <code@musicinmybrain.net> - 1.5.3-3
- Drop unnecessary weak dependency on python-pandas-datareader
- Backport proper pyarrow 10 and 11 support
* Thu Apr 13 2023 Benjamin A. Beasley <code@musicinmybrain.net> - 1.5.3-2
- Fix RHBZ#2171682 by backporting upstream PR#52150