def test_simple_rowtype():
df = DataFrame({'a': range(5), 'b': np.linspace(0, 1, 5)})
t1 = df.rowtype(pack=False)
assert (len(t1.fields) == 2)
assert (t1.names == ('a', 'b'))
t2 = df.rowtype(pack=True)
assert (t2.names == ('b', 'a'))
def test_repr():
df = DataFrame({
'a': np.arange(3),
'b': np.linspace(-1, 1, 3),
'c': np.ones(3, dtype=np.bool_),
'd': [1, 2, 3],
'e': [(1, ), (2, 3), (4, 5)],
'f': [[1, 2, 3], [4]]
})
df_copy = DataFrame(eval(repr(df)))
assert (len(df_copy) == len(df))
assert (set(df_copy.columns) == set(df.columns))
for key in df_copy:
for x, y in zip(df_copy[key], df[key]):
assert (x == y)
def test_column_attribute_access():
df = DataFrame({'a': range(5), 'b': np.linspace(0, 1, 5)})
assert (df.a is not None)
assert (df.b is not None)
with pytest.raises(AttributeError):
df.c
def test_pickle():
df1 = DataFrame()
df1['x'] = np.ones(10)
df1['y'] = np.arange(10)
pck = pickle.dumps(df1)
df2 = pickle.loads(pck)
assert ('x' in df2)
assert ('y' in df2)
assert (len(df2) == len(df1))
for x in df2.x:
assert (x == 1)
for i, y in enumerate(df2.y):
assert (y == i)
def test_drop_column():
df = DataFrame({'a': range(5), 'b': np.linspace(0, 1, 5)})
assert (list(df.drop(['a']).columns) == ['b'])
assert (list(df.drop(['b']).columns) == ['a'])
df_copy = df.copy()
df_copy.drop(['a'], inplace=True)
assert (list(df_copy.columns) == ['b'])
with pytest.raises(KeyError):
df.drop(['c'], inplace=True)
assert (len(df.drop(['c']).columns) == 2)
def test_masking():
df = DataFrame({'a': range(5), 'b': np.linspace(0, 1, 5)})
for mask in [
np.ones(5, dtype=np.bool_),
np.zeros(5, dtype=np.bool_),
np.array([True, True, False, False, False], dtype=np.bool_),
]:
df_masked = df[mask]
assert (isinstance(df_masked, DataFrame))
assert (len(df_masked) == np.count_nonzero(mask))
for c in df.columns:
assert (c in df_masked)
for c in df_masked:
assert (c in df)
def test_column_access():
df = DataFrame({'a': range(5), 'b': np.linspace(0, 1, 5)})
for invalid_key in [1, True, 1.5, None]:
with pytest.raises(TypeError):
df[invalid_key]
for nonexisting_col in ['d', 'e', 'f']:
with pytest.raises(KeyError):
df[nonexisting_col]
for sel in [['a'], ['a', 'b']]:
df_sel = df[sel]
assert (isinstance(df_sel, DataFrame))
assert (len(df_sel) == len(df))
assert (len(df_sel.columns) == len(sel))
with pytest.raises(KeyError):
df[['x', 'y']]
def test_row_iteration():
df = DataFrame({'a': range(5), 'b': np.linspace(0, 1, 5)})
for row in df.rows():
assert (len(row) == 2)
assert (row.a == row['a'])
assert (row.b == row['b'])
assert (row[0] == row['a'])
assert (row[1] == row['b'])
for i, row in df.iterrows():
assert (isinstance(i, int))
assert (len(row) == 2)
for tpl in df.itertuples():
assert (len(tpl) == 2)
for state in df.iterstates():
assert (len(state) == 2)
def test_column_iteration():
df = DataFrame({'a': range(5), 'b': np.linspace(0, 1, 5)})
columns = [col for col in df]
for c1, c2 in zip(columns, df.columns):
assert (c1 == c2)
def test_column_insertion():
df = DataFrame({'a': range(5)})
assert ('b' not in df)
df['b'] = np.linspace(0, 1, 5)
assert ('b' in df)
def test_dict_init():
df = DataFrame({'a': range(5)})
assert ('a' in df)
assert ('a' in df.columns)
assert (len(df) == 5)
assert (len(df.a) == 5)
def test_pickle_empty():
df1 = DataFrame()
pck = pickle.dumps(df1)
df2 = pickle.loads(pck)
assert (len(df1) == len(df2))
def test_rowtype(case):
data, dtype = case
df = DataFrame(data)
assert (df.rowtype(pack=False) == dtype)
assert (df.rowtype(pack=False).itemsize >= df.rowtype(pack=True).itemsize)
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import pickle
import numpy as np
import pytest
from graphdot.codegen.cpptool import cpptype
from graphdot.minipandas.dataframe import DataFrame
@pytest.mark.parametrize('df', [
DataFrame(),
DataFrame().copy(deep=True),
DataFrame().copy(deep=False),
])
def test_empty_df(df):
assert (len(df) == 0)
assert (df.columns == [])
assert (df.rowtype() == np.dtype([]))
assert (list(df.rows()) == [])
assert (list(df.iterrows()) == [])
assert (list(df.itertuples()) == [])
assert (list(df.iterstates()) == [])
pdf = df.to_pandas()
assert (len(pdf) == 0)
def test_dict_init():
df = DataFrame({'a': range(5)})
assert ('a' in df)
assert ('a' in df.columns)
assert (len(df) == 5)