Script 'mail_helper' called by obssrc
Hello community,
here is the log from the commit of package python-dask for openSUSE:Factory
checked in at 2023-03-28 17:50:14
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Comparing /work/SRC/openSUSE:Factory/python-dask (Old)
and /work/SRC/openSUSE:Factory/.python-dask.new.31432 (New)
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Package is "python-dask"
Tue Mar 28 17:50:14 2023 rev:64 rq:1074794 version:2023.3.2
Changes:
--------
--- /work/SRC/openSUSE:Factory/python-dask/python-dask.changes 2023-03-15
18:54:26.372361646 +0100
+++ /work/SRC/openSUSE:Factory/.python-dask.new.31432/python-dask.changes
2023-03-28 17:50:19.515252752 +0200
@@ -1,0 +2,38 @@
+Mon Mar 27 16:40:11 UTC 2023 - Ben Greiner <c...@bnavigator.de>
+
+- Update to 2023.3.2
+ ## Enhancements
+ * Deprecate observed=False for groupby with categoricals
+ (GH#10095) Irina Truong
+ * Deprecate axis= for some groupby operations (GH#10094) James
+ Bourbeau
+ * The axis keyword in DataFrame.rolling/Series.rolling is
+ deprecated (GH#10110) Irina Truong
+ * DataFrame._data deprecation in pandas (GH#10081) Irina Truong
+ * Use importlib_metadata backport to avoid CLI UserWarning
+ (GH#10070) Thomas Grainger
+ * Port option parsing logic from dask.dataframe.read_parquet to
+ to_parquet (GH#9981) Anton Loukianov
+ ## Bug Fixes
+ * Avoid using dd.shuffle in groupby-apply (GH#10043) Richard
+ (Rick) Zamora
+ * Enable null hive partitions with pyarrow parquet engine
+ (GH#10007) Richard (Rick) Zamora
+ * Support unknown shapes in *_like functions (GH#10064) Doug
+ Davis
+ ## Maintenance
+ * Restore Entrypoints compatibility (GH#10113) Jacob Tomlinson
+ * Allow pyarrow build to continue on failures (GH#10097) James
+ Bourbeau
+ * Fix test_set_index_on_empty with pyarrow strings active
+ (GH#10054) Irina Truong
+ * Temporarily skip pyarrow_compat tests with pandas 2.0
+ (GH#10063) James Bourbeau
+
+-------------------------------------------------------------------
+Sun Mar 26 17:13:15 UTC 2023 - Ben Greiner <c...@bnavigator.de>
+
+- Add dask-pr10042-parquetstats.patch gh#dask/dask#10042
+- Enable python311 build: numba is not a strict requirement
+
+-------------------------------------------------------------------
Old:
----
dask-2023.3.1.tar.gz
New:
----
dask-2023.3.2.tar.gz
dask-pr10042-parquetstats.patch
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Other differences:
------------------
++++++ python-dask.spec ++++++
--- /var/tmp/diff_new_pack.XnGtOj/_old 2023-03-28 17:50:20.127255705 +0200
+++ /var/tmp/diff_new_pack.XnGtOj/_new 2023-03-28 17:50:20.131255724 +0200
@@ -49,18 +49,18 @@
%if "%{flavor}" == ""
%bcond_with test
%endif
-# Numba is not ready for python 3.11 yet gh#numba/numba#8304
-%define skip_python311 1
Name: python-dask%{psuffix}
# ===> Note: python-dask MUST be updated in sync with python-distributed! <===
-Version: 2023.3.1
+Version: 2023.3.2
Release: 0
Summary: Minimal task scheduling abstraction
License: BSD-3-Clause
URL: https://dask.org
# SourceRepository: https://github.com/dask/dask
Source0:
https://files.pythonhosted.org/packages/source/d/dask/dask-%{version}.tar.gz
+# PATCH-FIX-UPSTREAM dask-pr10042-parquetstats.patch gh#dask/dask#10042
+Patch0: dask-pr10042-parquetstats.patch
BuildRequires: %{python_module base >= 3.8}
BuildRequires: %{python_module packaging >= 20.0}
BuildRequires: %{python_module pip}
@@ -72,6 +72,7 @@
Requires: python-click >= 7
Requires: python-cloudpickle >= 1.1.1
Requires: python-fsspec >= 0.6.0
+Requires: python-importlib-metadata >= 4.13.0
Requires: python-packaging >= 20.0
Requires: python-partd >= 1.2.0
Requires: python-toolz >= 0.8.2
@@ -116,13 +117,13 @@
BuildRequires: %{python_module matplotlib}
BuildRequires: %{python_module mimesis}
BuildRequires: %{python_module multipledispatch}
-BuildRequires: %{python_module numba}
+BuildRequires: %{python_module numba if %python-base < 3.11}
# snappy required for using fastparquet
BuildRequires: %{python_module python-snappy}
BuildRequires: %{python_module requests}
BuildRequires: %{python_module scikit-image}
BuildRequires: %{python_module scipy}
-BuildRequires: %{python_module sparse}
+BuildRequires: %{python_module sparse if %python-base < 3.11}
BuildRequires: %{python_module tables}
BuildRequires: %{python_module xarray if %python-base >= 3.9}
BuildRequires: %{python_module zarr}
@@ -153,9 +154,9 @@
Requires: %{name}-diagnostics = %{version}
Requires: %{name}-distributed = %{version}
Requires: %{name}-dot = %{version}
+Requires: python-lz4 >= 4.3.2
# Added to the [complete] extra in 2023.3.1, not available for TW yet
#Requires: python-pyarrow >= 7
-Requires: python-lz4 >= 4.3.2
Provides: %{name}-all = %{version}-%{release}
Obsoletes: %{name}-all < %{version}-%{release}
@@ -375,9 +376,6 @@
donttest="(test_datasets and test_deterministic)"
# upstreams test if their ci is up to date, irrelevant for obs
donttest+=" or test_development_guidelines_matches_ci"
-# requires otherwise optional pyarrow (not available on TW) --
https://github.com/dask/dask/issues/10042
-donttest+=" or (test_select_filtered_column and fastparquet)"
-donttest+=" or test_read_parquet_convert_string_fastparquet_warns"
if [[ $(getconf LONG_BIT) -eq 32 ]]; then
# https://github.com/dask/dask/issues/8620
donttest+=" or test_query_with_meta"
++++++ _multibuild ++++++
--- /var/tmp/diff_new_pack.XnGtOj/_old 2023-03-28 17:50:20.163255878 +0200
+++ /var/tmp/diff_new_pack.XnGtOj/_new 2023-03-28 17:50:20.167255898 +0200
@@ -2,6 +2,6 @@
<package>test-py38</package>
<package>test-py39</package>
<package>test-py310</package>
- <!-- package>test-py311</package -->
+ <package>test-py311</package>
</multibuild>
++++++ dask-2023.3.1.tar.gz -> dask-2023.3.2.tar.gz ++++++
/work/SRC/openSUSE:Factory/python-dask/dask-2023.3.1.tar.gz
/work/SRC/openSUSE:Factory/.python-dask.new.31432/dask-2023.3.2.tar.gz differ:
char 5, line 1
++++++ dask-pr10042-parquetstats.patch ++++++
diff --git a/dask/dataframe/io/parquet/core.py
b/dask/dataframe/io/parquet/core.py
index 07fa3a0e8f6..cc9ecaddb20 100644
--- a/dask/dataframe/io/parquet/core.py
+++ b/dask/dataframe/io/parquet/core.py
@@ -1375,14 +1375,20 @@ def apply_conjunction(parts, statistics, conjunction):
out_statistics.append(stats)
else:
if (
- operator != "is not"
- and min is None
- and max is None
- and null_count
+ # Must allow row-groups with "missing" stats
+ (min is None and max is None and not null_count)
+ # Check "is" and "is not" fiters first
or operator == "is"
and null_count
or operator == "is not"
and (not pd.isna(min) or not pd.isna(max))
+ # Allow all-null row-groups if not fitering out nulls
+ or operator != "is not"
+ and min is None
+ and max is None
+ and null_count
+ # Start conventional (non-null) fitering
+ # (main/max cannot be None for remaining checks)
or operator in ("==", "=")
and min <= value <= max
or operator == "!="
diff --git a/dask/dataframe/io/tests/test_parquet.py
b/dask/dataframe/io/tests/test_parquet.py
index f2403106eb9..f9c10d573dc 100644
--- a/dask/dataframe/io/tests/test_parquet.py
+++ b/dask/dataframe/io/tests/test_parquet.py
@@ -4718,10 +4718,28 @@ def test_fsspec_to_parquet_filesystem_option(tmp_path):
def test_select_filtered_column(tmp_path, engine):
df = pd.DataFrame({"a": range(10), "b": ["cat"] * 10})
path = tmp_path / "test_select_filtered_column.parquet"
- df.to_parquet(path, index=False)
+ stats = {"write_statistics" if engine == "pyarrow" else "stats": True}
+ df.to_parquet(path, engine=engine, index=False, **stats)
with pytest.warns(UserWarning, match="Sorted columns detected"):
ddf = dd.read_parquet(path, engine=engine, filters=[("b", "==",
"cat")])
+ assert_eq(df, ddf)
+
+ with pytest.warns(UserWarning, match="Sorted columns detected"):
+ ddf = dd.read_parquet(path, engine=engine, filters=[("b", "is not",
None)])
+ assert_eq(df, ddf)
+
+
+def test_select_filtered_column_no_stats(tmp_path, engine):
+ df = pd.DataFrame({"a": range(10), "b": ["cat"] * 10})
+ path = tmp_path / "test_select_filtered_column_no_stats.parquet"
+ stats = {"write_statistics" if engine == "pyarrow" else "stats": False}
+ df.to_parquet(path, engine=engine, **stats)
+
+ ddf = dd.read_parquet(path, engine=engine, filters=[("b", "==", "cat")])
+ assert_eq(df, ddf)
+
+ ddf = dd.read_parquet(path, engine=engine, filters=[("b", "is not", None)])
assert_eq(df, ddf)
@@ -4793,6 +4811,7 @@ def
test_read_parquet_convert_string_nullable_mapper(tmp_path, engine):
assert_eq(ddf, expected)
+@PYARROW_MARK # We get an error instead of a warning without pyarrow
@FASTPARQUET_MARK
@pytest.mark.skipif(
not PANDAS_GT_200, reason="dataframe.convert_string requires pandas>=2.0"
