jorisvandenbossche commented on pull request #7891:
URL: https://github.com/apache/arrow/pull/7891#issuecomment-670478968
cc @fjetter @xhochy in case if are familiar with those failures
There are several failures like this (related to exit codes):
```
_________________________________ test_simple
__________________________________
cli = <function cli.<locals>._cli at 0x7ff27b5fed40>
built_cube = Cube(dimension_columns=('x', 'y'), partition_columns=('p',
'q'), uuid_prefix='my_cube', seed_dataset='source',
index_columns=frozenset({'i2', 'i1'}))
mock_prompt = <function mock_prompt.<locals>._f at 0x7ff27b5b6830>
df_complete = i1 i2 p ...
v2 x y
0 True 2018-01-01 0 ... ... 8, 9] 1 0
3 False 2019-01-01 0 ... [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,
13,... 1 1
[4 rows x 9 columns]
skv = None
def test_simple(cli, built_cube, mock_prompt, df_complete, skv):
mock_prompt(["", ""]) # conditions # payload
result = cli("--store=cubes", "my_cube", "query", input="df\n")
str_df = str(df_complete.loc[:, ["p", "q", "x",
"y"]].reset_index(drop=True))
> assert result.exit_code == 1
E assert 2 == 1
E + where 2 = <Result SystemExit(2)>.exit_code
/kartothek/tests/cli/test_query.py:64: AssertionError
```
and there are a few failures like
```
________ test_predicate_pushdown_null_col[True-serialiser4-df3-value3]
_________
store = <conftest.FakeStore object at 0x7ff242feef50>, df = t
0 NaT
value = Timestamp('2017-01-01 00:00:00'), predicate_pushdown_to_io = True
serialiser = CsvSerializer(compress=True)
@pytest.mark.parametrize(
"df,value",
[
(pd.DataFrame({"u": pd.Series([None], dtype=object)}), "foo"),
(pd.DataFrame({"b": pd.Series([None], dtype=object)}), b"foo"),
(pd.DataFrame({"f": pd.Series([np.nan], dtype=float)}), 1.2),
(
pd.DataFrame({"t": pd.Series([pd.NaT],
dtype="datetime64[ns]")}),
pd.Timestamp("2017"),
),
],
)
@predicate_serialisers
@pytest.mark.parametrize("predicate_pushdown_to_io", [True, False])
def test_predicate_pushdown_null_col(
store, df, value, predicate_pushdown_to_io, serialiser
):
key = serialiser.store(store, "prefix", df)
expected = df.iloc[[]].copy()
predicates = [[(df.columns[0], "==", value)]]
result = serialiser.restore_dataframe(
store,
key,
predicate_pushdown_to_io=predicate_pushdown_to_io,
predicates=predicates,
)
pdt.assert_frame_equal(
result.reset_index(drop=True),
expected.reset_index(drop=True),
> check_dtype=serialiser.type_stable,
)
tests/serialization/test_dataframe.py:529:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _
left = array([], dtype=float64)
right = <DatetimeArray>
[]
Length: 0, dtype: datetime64[ns], check_dtype = True
index_values = array([], dtype=int64)
check_less_precise = <object object at 0x7ff2986040f0>, check_exact = False
rtol = 1e-05, atol = 1e-08
> assert isinstance(left, ExtensionArray), "left is not an
ExtensionArray"
E AssertionError: left is not an ExtensionArray
/opt/conda/envs/arrow/lib/python3.7/site-packages/pandas/_testing.py:1175:
AssertionError
```
Since this is related to a pandas testing utility, does kartothek already
support pandas 1.1?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
