def test_spatial_heterogeneous():
recs = {
'p': [neo4j.spatial.Point([1, 2, 3, 4]), 1],
'c': [neo4j.spatial.CartesianPoint([1, 2]), 1],
'c2': [neo4j.spatial.CartesianPoint([1, 2, 3]), 1],
'w': [neo4j.spatial.WGS84Point([4, 5]), 1]
}
df = pd.DataFrame(recs)
df2 = neo_df_to_pd_df(df)
assert df2.dtypes.to_dict() == {
'p': 'object',
'c': 'object',
'c2': 'object',
'w': 'object',
}
d = df2.to_dict(orient='records')[0]
assert d == {
'p': 'POINT(1 2 3 4)',
'c': 'POINT(1 2)',
'c2': 'POINT(1 2 3)',
'w': 'POINT(4 5)',
}
assert df2.to_dict(orient='records')[1] == {
'p': 1,
'c': 1,
'c2': 1,
'w': 1,
}
with pytest.raises(pa.lib.ArrowTypeError):
pa.Table.from_pandas(df2)
def test_dates_homogeneous_na():
recs = {
'd': [neo4j.time.Date(2020, 10, 20), None],
'dt': [neo4j.time.DateTime(2020, 10, 20, 3, 4, 5), None],
't': [neo4j.time.Time(10, 20, 30), None],
'dur': [neo4j.time.Duration(1, 3, 2), None]
}
df = pd.DataFrame(recs)
df2 = neo_df_to_pd_df(df)
assert df2.dtypes.to_dict() == {
'd': 'datetime64[ns]',
'dt': 'datetime64[ns]',
't': 'timedelta64[ns]',
'dur': 'object'
}
d = df2.to_dict(orient='records')[0]
assert d == {
'd': dt.datetime(2020, 10, 20),
'dt': dt.datetime(2020, 10, 20, 3, 4, 5),
't': pd.to_timedelta('10:20:30'),
'dur': 'P1Y3M14D'
}
assert df2.to_dict(orient='records')[1] == {
'd': pd.NaT,
'dt': pd.NaT,
't': pd.NaT,
'dur': None
}
pa.Table.from_pandas(df2)
def test_spatial_homogenous_na():
recs = {
'p': [neo4j.spatial.Point([1, 2, 3, 4]), None],
'c': [neo4j.spatial.CartesianPoint([1, 2]), None],
'c2': [neo4j.spatial.CartesianPoint([1, 2, 3]), None],
'w': [neo4j.spatial.WGS84Point([4, 5]), None]
}
df = pd.DataFrame(recs)
df2 = neo_df_to_pd_df(df)
assert df2.dtypes.to_dict() == {
'p': 'object',
'p_srid': 'object',
'c': 'object',
'c_x': 'float64',
'c_y': 'float64',
'c_srid': 'float64',
'c2': 'object',
'c2_x': 'float64',
'c2_y': 'float64',
'c2_z': 'float64',
'c2_srid': 'float64',
'w': 'object',
'w_x': 'float64',
'w_y': 'float64',
'w_longitude': 'float64',
'w_latitude': 'float64',
'w_srid': 'float64',
}
d = df2.to_dict(orient='records')[0]
assert d == {
'p': 'POINT(1 2 3 4)',
'p_srid': None,
'c': 'POINT(1 2)',
'c_x': 1,
'c_y': 2,
'c_srid': 7203,
'c2': 'POINT(1 2 3)',
'c2_x': 1,
'c2_y': 2,
'c2_z': 3,
'c2_srid': 9157,
'w': 'POINT(4 5)',
'w_x': 4,
'w_y': 5,
'w_longitude': 4,
'w_latitude': 5,
'w_srid': 4326
}
d2 = df2.to_dict(orient='records')[1]
assert d2['p_srid'] is None
for k in d2.keys():
if k not in ['p', 'c', 'c2', 'w', 'p_srid']:
assert pd.isna(d2[k])
pa.Table.from_pandas(df2)
def test_spatial_homogenous():
rec = {
'p': neo4j.spatial.Point([1,2,3,4]),
'c': neo4j.spatial.CartesianPoint([1,2]),
'c2': neo4j.spatial.CartesianPoint([1,2,3]),
'w': neo4j.spatial.WGS84Point([4,5])
}
df = pd.DataFrame([rec])
df2 = neo_df_to_pd_df(df)
assert df2.dtypes.to_dict() == {
'p': 'object',
'p_srid': 'object',
'c': 'object',
'c_x': 'int64',
'c_y': 'int64',
'c_srid': 'int64',
'c2': 'object',
'c2_x': 'int64',
'c2_y': 'int64',
'c2_z': 'int64',
'c2_srid': 'int64',
'w': 'object',
'w_x': 'int64',
'w_y': 'int64',
'w_longitude': 'int64',
'w_latitude': 'int64',
'w_srid': 'int64',
}
d = df2.to_dict(orient='record')[0]
assert d == {
'p': 'POINT(1 2 3 4)',
'p_srid': None,
'c': 'POINT(1 2)',
'c_x': 1,
'c_y': 2,
'c_srid': 7203,
'c2': 'POINT(1 2 3)',
'c2_x': 1,
'c2_y': 2,
'c2_z': 3,
'c2_srid': 9157,
'w': 'POINT(4 5)',
'w_x': 4,
'w_y': 5,
'w_longitude': 4,
'w_latitude': 5,
'w_srid': 4326
}
pa.Table.from_pandas(df2)
def test_neo_df_to_pd_df_basics():
rec = {
'x': 1,
'b': True,
's': 'abc',
'a': [1,2,3],
'd': {'r': 'v'},
'mt': None
}
df = pd.DataFrame([rec])
df2 = neo_df_to_pd_df(df)
assert df2.dtypes.to_dict() == {
'x': 'int',
'b': 'bool',
's': 'object',
'a': 'object',
'd': 'object',
'mt': 'object'
}
d = df2.to_dict(orient='record')[0]
assert d == rec
pa.Table.from_pandas(df2)
def test_neo_df_to_pd_df_basics_na():
recs = {
'x': [1, None],
'b': [True, None],
's': ['abc', None],
'a': [[1,2,3], None],
'd': [{'r': 'v'}, None],
'mt': [None, None]
}
df = pd.DataFrame(recs)
df2 = neo_df_to_pd_df(df)
assert df2.dtypes.to_dict() == {
'x': 'float64',
'b': 'object',
's': 'object',
'a': 'object',
'd': 'object',
'mt': 'object'
}
d = df2.to_dict(orient='record')[0]
assert d == {k: recs[k][0] for k in recs.keys()}
pa.Table.from_pandas(df2)
def test_dates_heterogeneous():
recs = {
'd': [neo4j.time.Date(2020, 10, 20), 1],
'dt': [neo4j.time.DateTime(2020, 10, 20, 3, 4, 5), 1],
't': [neo4j.time.Time(10, 20, 30), 1],
'dur': [neo4j.time.Duration(1, 3, 2), 1]
}
df = pd.DataFrame(recs)
df2 = neo_df_to_pd_df(df)
assert df.dtypes.to_dict() == {
'd': 'object',
'dt': 'object',
't': 'object',
'dur': 'object'
}
assert df2.dtypes.to_dict() == {
'd': 'object',
'dt': 'object',
't': 'object',
'dur': 'object'
}
d = df2.to_dict(orient='records')[0]
assert d == {
'd': dt.datetime(2020, 10, 20),
'dt': dt.datetime(2020, 10, 20, 3, 4, 5),
't': pd.to_timedelta('10:20:30'),
'dur': 'P1Y3M14D'
}
assert df2.to_dict(orient='records')[1] == {
'd': 1,
'dt': 1,
't': 1,
'dur': 1
}
with pytest.raises(pa.lib.ArrowTypeError):
pa.Table.from_pandas(df2)
