def test_parquet(self):
with tmpfile(".parquet") as path:
df = UntypedDf(sample_data())
df.to_parquet(path)
df2 = UntypedDf.read_parquet(path)
assert list(df2.index.names) == [None]
assert set(df2.columns) == {"abc", "123", "xyz"}
def test_of(self):
expected = [[1, 2, 3], [4, 5, 6]]
df = UntypedDf.convert(pd.DataFrame(sample_data()))
assert df.to_numpy().tolist() == expected
df = UntypedDf.of(pd.DataFrame(sample_data()))
assert df.to_numpy().tolist() == expected
df = UntypedDf.of(sample_data())
assert df.to_numpy().tolist() == expected
def test_concat_no_preserve_attrs(self):
df0 = Col1([pd.Series(dict(abc="hippo"))])
df1 = df0.set_attrs(animal="fishies")
df2 = df0.set_attrs(animal="hippos")
df = UntypedDf.of([df1, df2])
assert df0.attrs == {}
assert len(df) == len(df1) + len(df2) > 0
assert df.attrs == {}
def test_read_properties(self):
data = r"""
[section]
# a comment
! another comment
k\:e\\y = v:a\\lue
"""
s = StringIO(data)
df = UntypedDf.read_properties(s)
assert df.column_names() == ["key", "value"]
assert df.values.tolist() == [[r"section.k:e\y", r"v:a\lue"]]
data: str = df.to_properties()
lines = [s.strip() for s in data.splitlines()]
assert "[section]" in lines
assert r"k\:e\\y = v:a\\lue" in lines
s = StringIO(data)
df2 = UntypedDf.read_properties(s)
assert df2.values.tolist() == df.values.tolist()
def test_st(self):
df = UntypedDf().convert(pd.DataFrame(sample_data()))
assert len(df[df["xyz"] == 6]) == 1
assert len(df.st(df["xyz"] == 6)) == 1
assert len(df.st(xyz=6)) == 1
assert len(df.st(df["xyz"] == 6, df["xyz"] == 6)) == 1
assert len(df.st(df["xyz"] == 6, df["xyz"] == 3)) == 0
assert len(df.st(df["xyz"] == 6, xyz=6)) == 1
assert len(df.st(df["xyz"] == 6, xyz=2)) == 0
assert len(df.st(xyz=1, abc=1)) == 0
assert len(df.st(df["xyz"] == 6, df["xyz"] == 1, xyz=6)) == 0
def test_read_ini(self):
data = """
[section]
; a comment
key = value
"""
s = StringIO(data)
df = UntypedDf.read_ini(s)
assert df.column_names() == ["key", "value"]
assert df.values.tolist() == [["section.key", "value"]]
def test_read_toml(self):
data = """
[[row]]
# a comment
key = "value"
"""
s = StringIO(data)
df = UntypedDf.read_toml(s)
assert df.column_names() == ["key"]
assert df.values.tolist() == [["value"]]
def test_read_toml_jagged(self):
data = """
[[row]]
key = "value1"
[[row]]
key = "value2"
kitten = "elephant"
cuteness = 10.3
"""
s = StringIO(data)
df = UntypedDf.read_toml(s)
assert df.column_names() == ["key", "kitten", "cuteness"]
xx = df.fillna(0).values.tolist()
assert xx == [["value1", 0, 0], ["value2", "elephant", 10.3]]
def test_untyped_read_write_csv(self):
with tmpfile(".csv") as path:
for indices in [None, "abc", ["abc", "xyz"]]:
df = UntypedDf(sample_data())
if indices is not None:
df = df.set_index(indices)
df.to_csv(path)
df2 = UntypedDf.read_csv(path)
assert list(df2.index.names) == [None]
assert set(df2.columns) == {"abc", "123", "xyz"}
def test_html_empty(self):
with tmpfile(".html") as path:
path.write_text("<html></html>", encoding="utf8")
with pytest.raises(NoValueError):
UntypedDf.read_html(path)
def test_set_index(self):
df = UntypedDf.convert(pd.DataFrame(sample_data()).set_index("abc"))
assert df.set_index([]).index_names() == []
assert df.set_index([], append=True).index_names() == ["abc"]
with pytest.raises(UnsupportedOperationError):
df.set_index([], inplace=True)
def test_only(self):
df = UntypedDf().convert(pd.DataFrame(sample_data_2()))
with pytest.raises(ValueError):
df.only("multi", exclude_na=True)
with pytest.raises(KeyError):
df.only("abc")
assert df.only("only") == 1
assert df.only("only", exclude_na=True) == 1
assert pd.isna(df.only("none"))
with pytest.raises(ValueError):
df.only("none", exclude_na=True)
def test_of_concat(self):
df1 = UntypedDf.of(pd.DataFrame(sample_data()))
df2 = UntypedDf.of(pd.DataFrame(sample_data()))
df = UntypedDf.of([df1, df2])
assert len(df) == len(df1) + len(df2) > 0
def test_set_attrs(self):
df = UntypedDf.convert(pd.DataFrame(sample_data()))
df2 = df.set_attrs(animal="fishies")
assert df2.attrs == dict(animal="fishies")
assert df.attrs == {}
def test_iter_rc(self):
df = UntypedDf.convert(pd.DataFrame(sample_data()))
expected = [((0, 0), 1), ((0, 1), 2), ((0, 2), 3), ((1, 0), 4),
((1, 1), 5), ((1, 2), 6)]
assert list(df.iter_row_col()) == expected
def test_records(self):
df = UntypedDf(sample_data())
records = df.to_records()
df2 = UntypedDf.from_records(records)
assert isinstance(df2, UntypedDf)
def test_write_passing_index(self):
with tmpfile(".csv") as path:
df = Trivial(sample_data())
df.to_csv(path, index=["abc"]) # fine
df = UntypedDf(sample_data())
df.to_csv(path, index=["abc"]) # calls super immediately
def test_html_invalid(self):
with tmpfile(".html") as path:
path.write_text("", encoding="utf8")
with pytest.raises(XMLSyntaxError):
UntypedDf.read_html(path)
def test_pretty(self):
assert (UntypedDf()._repr_html_().startswith(
"<strong>UntypedDf: 0 rows × 0 columns</strong>"))
def test_pretty(self):
assert (UntypedDf()._repr_html_().startswith(
"<strong>UntypedDf: 0 rows × 0 columns</strong>"))
df = UntypedDf(sample_data()).set_index(["abc", "123"])
assert (UntypedDf(df)._repr_html_().startswith(
"<strong>UntypedDf: 2 rows × 1 columns, 2 index columns</strong>"))