def test_tail(self, df_small):
actual = df_small.tail(2)
data = [4, 5]
expected = DataFrame(
OrderedDict(
(('a', np.array(data, dtype=np.float32)),
('b', np.array(data)), ('c',
np.array([' dC ', 'secrETariat'],
dtype=np.dtype('|S11'))))),
Index(np.array([3, 4])))
assert_dataframe_equal(actual, expected)
def test_filter(self, df_small):
actual = df_small[Series(np.array([False, True, True, False, False]))]
data = [2, 3]
expected = DataFrame(
OrderedDict(
(('a', np.array(data,
dtype=np.float32)), ('b', np.array(data)),
('c', np.array(['Abc', 'goosfraba'],
dtype=np.dtype('|S11'))))),
Index(np.array([1, 2])))
assert_dataframe_equal(actual, expected)
def test_merge_sorted_unique_multi_on_inner(self, df1, df2):
actual = df1.merge(df2, on=['a', 'b'], is_on_sorted=True)
expected = DataFrame(
OrderedDict(
(('index', np.array([5])), ('d',
np.array(['def'],
dtype=np.dtype('|S4'))),
('c', np.array([5])))),
MultiIndex(
[np.array([3]), np.array([4], dtype=np.float32)], ['a', 'b']))
assert_dataframe_equal(actual, expected)
def test_merge_sorted_unique_single_on_inner(self, df1, df2):
actual = df1.merge(df2, on='a')
expected = DataFrame(
OrderedDict(
(('index', np.array([3, 5])), ('b_x',
np.array([2, 4],
dtype=np.float32)),
('d', np.array(['abc', 'def'], dtype=np.dtype('|S4'))),
('b_y', Series(np.arange(3, 5, dtype=np.float32))),
('c', np.arange(4, 6)))),
Index(np.array([1, 3]), np.dtype(np.int64), 'a'))
assert_dataframe_equal(actual, expected)
def test_iloc_indices(self, df_small):
indices = Series(np.array([0, 2, 3]))
actual = df_small.iloc[indices]
data = [1, 3, 4]
expected = DataFrame(
OrderedDict(
(('a', np.array(data,
dtype=np.float32)), ('b', np.array(data)),
('c',
np.array(['a', 'goosfraba', ' dC '],
dtype=np.dtype('|S11'))))),
Index(np.array([0, 2, 3])))
assert_dataframe_equal(actual, expected)
def test_merge_sorted_unique_single_on_right(self, df1, df2):
actual = df1.merge(df2, on='a', how='right')
expected = DataFrame(
OrderedDict(
(('index', np.array([3, 5,
-999])), ('b_x',
np.array([2, 4, -999.],
dtype=np.float32)),
('d', np.array(['abc', 'def', 'efgh'],
dtype=np.dtype('|S4'))),
('b_y', Series(np.array([3., 4., 5.], dtype=np.float32))),
('c', np.array([4, 5, 6])))),
Index(np.array([1, 3, 5]), np.dtype(np.int64), 'a'))
assert_dataframe_equal(actual, expected)
def test_merge_sorted_unique_single_on_left(self, df1, df2):
actual = df1.merge(df2, on='a', how='left')
expected = DataFrame(
OrderedDict(
(('index', np.arange(2, 7)), ('b_x',
np.arange(1, 6,
dtype=np.float32)),
('d',
np.array(['None', 'abc', 'None', 'def', 'None'],
dtype=np.bytes_)),
('b_y',
Series(
np.array([-999., 3., -999., 4., -999.],
dtype=np.float32))),
('c', np.array([-999, 4, -999, 5, -999])))),
Index(np.arange(5), np.dtype(np.int64), 'a'))
assert_dataframe_equal(actual, expected)
def df_empty():
return DataFrame()
def df_small(data_f32, series_i64, series_str, index_i64):
return DataFrame(OrderedDict((('a', data_f32), ('b', series_i64), ('c', series_str))), index_i64)
def df_dupl(series_i64, index_i64):
return DataFrame(OrderedDict((('a', np.array([0, 1, 1, 2, 3], dtype=np.float32)),
('b', [4, 5, 5, 6, 6]),
('c', series_i64))),
index_i64)
def test_astype_dtype(self, df1, data_i64, index_i64_2):
actual = df1.astype(np.dtype(np.int64))
expected = DataFrame(
OrderedDict((('a', np.arange(5)), ('b', data_i64))), index_i64_2)
assert_dataframe_equal(actual, expected)
def test_set_index(self, df_small, data_f32, data_i64, data_str):
actual = df_small.set_index('b')
expected = DataFrame(OrderedDict((('a', data_f32), ('c', data_str))),
Index(data_i64, np.dtype(np.int64), 'b'))
assert_dataframe_equal(actual, expected)
def test_drop_multi(self, df_small, series_i64, index_i64):
actual = df_small.drop(['a', 'c'])
expected = DataFrame({'b': series_i64})
assert_dataframe_equal(actual, expected)
def test_drop_single(self, df_small, data_f32, series_i64, index_i64):
actual = df_small.drop('c')
expected = DataFrame(OrderedDict((('a', data_f32), ('b', series_i64))),
index_i64)
assert_dataframe_equal(actual, expected)
def df1(data_f32, index_i64_2):
return DataFrame(OrderedDict((('a', Series(np.arange(5))), ('b', data_f32))), index_i64_2)
def test_init_list(self):
data = [1, 2, 3]
actual = DataFrame({'a': data})
expected = DataFrame({'a': np.array(data)})
assert_dataframe_equal(actual, expected)
def df2():
return DataFrame(OrderedDict((('b', np.arange(3, 6, dtype=np.float32)), ('c', np.arange(4, 7)))),
MultiIndex([np.array([1, 3, 5]),
Index(np.array(['abc', 'def', 'efgh'], dtype=np.bytes_))],
['a', 'd']))
def test_empty_dataframe_init():
assert_dataframe_equal(DataFrame(), DataFrame({}, Index(np.empty(0, dtype=np.int64))))