wassname/ray
1
0
Fork 0
mirror of https://github.com/wassname/ray.git synced 2026-06-29 17:04:56 +08:00
Code Issues Packages Projects Releases Wiki Activity

[DataFrame] Adding insert, set_axis, set_index, reset_index and tests (#1603)

Browse Source
This commit is contained in:
Devin Petersohn
2018-02-26 08:58:15 -08:00
committed by Robert Nishihara
parent c2ad800cbf
commit 1fa59f1887
2 changed files with 384 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/ray/dataframe/test/test_dataframe.py
+119 -24
View File
@@ -216,6 +216,29 @@ def test_int_dataframe():
test_notna(ray_df, pandas_df)
test_notnull(ray_df, pandas_df)
labels = ['a', 'b', 'c', 'd']
test_set_axis(ray_df, pandas_df, labels, 0)
test_set_axis(ray_df, pandas_df, labels, 'rows')
labels.append('e')
test_set_axis(ray_df, pandas_df, labels, 1)
test_set_axis(ray_df, pandas_df, labels, 'columns')
for key in keys:
test_set_index(ray_df, pandas_df, key)
test_reset_index(ray_df, pandas_df)
test_reset_index(ray_df, pandas_df, inplace=True)
for key in keys:
test___contains__(ray_df, key, True)
test___contains__(ray_df, "Not Exists", False)
for key in keys:
test_insert(ray_df, pandas_df, 0, "New Column", ray_df[key])
test_insert(ray_df, pandas_df, 0, "New Column", pandas_df[key])
test_insert(ray_df, pandas_df, 1, "New Column", ray_df[key])
test_insert(ray_df, pandas_df, 4, "New Column", ray_df[key])
def test_float_dataframe():
@@ -288,6 +311,26 @@ def test_float_dataframe():
test_iteritems(ray_df, pandas_df)
test_itertuples(ray_df, pandas_df)
labels = ['a', 'b', 'c', 'd']
test_set_axis(ray_df, pandas_df, labels, 0)
test_set_axis(ray_df, pandas_df, labels, 'rows')
labels.append('e')
test_set_axis(ray_df, pandas_df, labels, 1)
test_set_axis(ray_df, pandas_df, labels, 'columns')
for key in keys:
test_set_index(ray_df, pandas_df, key)
test_set_index(ray_df, pandas_df, key, inplace=True)
test_reset_index(ray_df, pandas_df)
test_reset_index(ray_df, pandas_df, inplace=True)
for key in keys:
test_insert(ray_df, pandas_df, 0, "New Column", ray_df[key])
test_insert(ray_df, pandas_df, 0, "New Column", pandas_df[key])
test_insert(ray_df, pandas_df, 1, "New Column", ray_df[key])
test_insert(ray_df, pandas_df, 4, "New Column", ray_df[key])
def test_mixed_dtype_dataframe():
pandas_df = pd.DataFrame({
@@ -365,6 +408,25 @@ def test_mixed_dtype_dataframe():
test_iteritems(ray_df, pandas_df)
test_itertuples(ray_df, pandas_df)
labels = ['a', 'b', 'c', 'd']
test_set_axis(ray_df, pandas_df, labels, 0)
test_set_axis(ray_df, pandas_df, labels, 'rows')
test_set_axis(ray_df, pandas_df, labels, 1)
test_set_axis(ray_df, pandas_df, labels, 'columns')
for key in keys:
test_set_index(ray_df, pandas_df, key)
test_set_index(ray_df, pandas_df, key, inplace=True)
test_reset_index(ray_df, pandas_df)
test_reset_index(ray_df, pandas_df, inplace=True)
for key in keys:
test_insert(ray_df, pandas_df, 0, "New Column", ray_df[key])
test_insert(ray_df, pandas_df, 0, "New Column", pandas_df[key])
test_insert(ray_df, pandas_df, 1, "New Column", ray_df[key])
test_insert(ray_df, pandas_df, 4, "New Column", ray_df[key])
def test_nan_dataframe():
pandas_df = pd.DataFrame({
@@ -435,6 +497,25 @@ def test_nan_dataframe():
test_iteritems(ray_df, pandas_df)
test_itertuples(ray_df, pandas_df)
labels = ['a', 'b', 'c', 'd']
test_set_axis(ray_df, pandas_df, labels, 0)
test_set_axis(ray_df, pandas_df, labels, 'rows')
test_set_axis(ray_df, pandas_df, labels, 1)
test_set_axis(ray_df, pandas_df, labels, 'columns')
for key in keys:
test_set_index(ray_df, pandas_df, key)
test_set_index(ray_df, pandas_df, key, inplace=True)
test_reset_index(ray_df, pandas_df)
test_reset_index(ray_df, pandas_df, inplace=True)
for key in keys:
test_insert(ray_df, pandas_df, 0, "New Column", ray_df[key])
test_insert(ray_df, pandas_df, 0, "New Column", pandas_df[key])
test_insert(ray_df, pandas_df, 1, "New Column", ray_df[key])
test_insert(ray_df, pandas_df, 4, "New Column", ray_df[key])
def test_add():
ray_df = create_test_dataframe()
@@ -902,11 +983,13 @@ def test_info():
ray_df.info()
def test_insert():
ray_df = create_test_dataframe()
@pytest.fixture
def test_insert(ray_df, pandas_df, loc, column, value):
ray_df_cp = ray_df.copy()
pd_df_cp = pandas_df.copy()
with pytest.raises(NotImplementedError):
ray_df.insert(None, None, None)
ray_df_cp.insert(loc, column, value)
pd_df_cp.insert(loc, column, value)
def test_interpolate():
@@ -1307,11 +1390,19 @@ def test_resample():
ray_df.resample(None)
def test_reset_index():
ray_df = create_test_dataframe()
with pytest.raises(NotImplementedError):
ray_df.reset_index()
@pytest.fixture
def test_reset_index(ray_df, pandas_df, inplace=False):
if not inplace:
print(rdf.to_pandas(ray_df.reset_index(inplace=inplace)).index)
print(pandas_df.reset_index(inplace=inplace))
assert rdf.to_pandas(ray_df.reset_index(inplace=inplace)).equals(
pandas_df.reset_index(inplace=inplace))
else:
ray_df_cp = ray_df.copy()
pd_df_cp = pandas_df.copy()
ray_df_cp.reset_index(inplace=inplace)
pd_df_cp.reset_index(inplace=inplace)
assert rdf.to_pandas(ray_df_cp).equals(pd_df_cp)
def test_rfloordiv():
@@ -1397,18 +1488,23 @@ def test_sem():
ray_df.sem()
def test_set_axis():
ray_df = create_test_dataframe()
with pytest.raises(NotImplementedError):
ray_df.set_axis(None)
@pytest.fixture
def test_set_axis(ray_df, pandas_df, label, axis):
assert rdf.to_pandas(ray_df.set_axis(label, axis, inplace=False)).equals(
pandas_df.set_axis(label, axis, inplace=False))
def test_set_index():
ray_df = create_test_dataframe()
with pytest.raises(NotImplementedError):
ray_df.set_index(None)
@pytest.fixture
def test_set_index(ray_df, pandas_df, keys, inplace=False):
if not inplace:
assert rdf.to_pandas(ray_df.set_index(keys)).equals(
pandas_df.set_index(keys))
else:
ray_df_cp = ray_df.copy()
pd_df_cp = pandas_df.copy()
ray_df_cp.set_index(keys, inplace=inplace)
pd_df_cp.set_index(keys, inplace=inplace)
assert rdf.to_pandas(ray_df_cp).equals(pd_df_cp)
def test_set_value():
@@ -1817,11 +1913,10 @@ def test___iter__():
ray_df.__iter__()
def test___contains__():
ray_df = create_test_dataframe()
with pytest.raises(NotImplementedError):
ray_df.__contains__(None)
@pytest.fixture
def test___contains__(ray_df, key, result):
assert result == ray_df.__contains__(key)
assert result == (key in ray_df)
def test___nonzero__():