Skip to content

Commit a721f1e

Browse files
xhochyMeeseeksDev[bot]
authored and
MeeseeksDev[bot]
committed
Backport PR pandas-dev#28357: Change conda channel order for Windows builds
1 parent 69e8bad commit a721f1e

File tree

2 files changed

+11
-3
lines changed

2 files changed

+11
-3
lines changed

ci/deps/azure-windows-36.yaml

+1-3
Original file line numberDiff line numberDiff line change
@@ -1,17 +1,15 @@
11
name: pandas-dev
22
channels:
3-
- defaults
43
- conda-forge
4+
- defaults
55
dependencies:
66
- blosc
77
- bottleneck
8-
- boost-cpp<1.67
98
- fastparquet>=0.2.1
109
- matplotlib=3.0.2
1110
- numexpr
1211
- numpy=1.15.*
1312
- openpyxl
14-
- parquet-cpp
1513
- pyarrow
1614
- pytables
1715
- python-dateutil

pandas/tests/io/test_parquet.py

+10
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,6 @@
11
""" test parquet compat """
22
import datetime
3+
from distutils.version import LooseVersion
34
import os
45
from warnings import catch_warnings
56

@@ -238,6 +239,15 @@ def test_cross_engine_pa_fp(df_cross_compat, pa, fp):
238239
def test_cross_engine_fp_pa(df_cross_compat, pa, fp):
239240
# cross-compat with differing reading/writing engines
240241

242+
if (
243+
LooseVersion(pyarrow.__version__) < "0.15"
244+
and LooseVersion(pyarrow.__version__) >= "0.13"
245+
):
246+
pytest.xfail(
247+
"Reading fastparquet with pyarrow in 0.14 fails: "
248+
"https://issues.apache.org/jira/browse/ARROW-6492"
249+
)
250+
241251
df = df_cross_compat
242252
with tm.ensure_clean() as path:
243253
df.to_parquet(path, engine=fp, compression=None)

0 commit comments

Comments
 (0)