def test_column_regex_matching(column_name_regex, expected_matches, error):
"""
Column regex pattern matching should yield correct matches and raise
expected errors.
"""
columns = pd.MultiIndex.from_tuples((
("foo_1", "biz_1"),
("foo_2", "baz_1"),
("foo_3", "baz_2"),
("bar_1", "biz_2"),
("bar_2", "biz_3"),
("bar_3", "biz_3"),
))
column_schema = Column(
Int,
Check(lambda s: s >= 0),
name=column_name_regex,
regex=True,
)
if error is not None:
with pytest.raises(error):
column_schema.get_regex_columns(columns)
else:
matched_columns = column_schema.get_regex_columns(columns)
assert expected_matches == matched_columns.tolist()
def test_column_regex_matching_non_str_types(
column_name_regex: str, expected_matches: List
) -> None:
"""Non-string column names should be cast into str for regex matching."""
columns = pd.Index([1, 2.2, 3.1415, -1, -3.6, pd.Timestamp("2018/01/01")])
column_schema = Column(name=column_name_regex, regex=True)
matched_columns = column_schema.get_regex_columns(columns)
assert expected_matches == matched_columns.tolist()
def test_column_regex_matching_non_str_types_multiindex(
column_name_regex: Tuple[str, str], expected_matches: List[Tuple[Any, Any]]
) -> None:
"""
Non-string column names should be cast into str for regex matching in
MultiIndex column case.
"""
columns = pd.MultiIndex.from_tuples(
(
("foo", 1),
("foo", pd.Timestamp("2018/01/01")),
(1, 2.2),
(3.14, -1),
)
)
column_schema = Column(name=column_name_regex, regex=True)
matched_columns = column_schema.get_regex_columns(columns)
assert expected_matches == matched_columns.tolist()
