def test_tonumpy_with_NAs2(numpy):
src = [[2.3, 11.89, None, math.inf], [4, None, None, -12]]
d0 = dt.Frame(src)
a0 = d0.to_numpy()
assert a0.T.tolist() == src
def test_rename_bad2():
d0 = dt.Frame([[1], [2], ["hello"]], names=("a", "b", "c"))
with pytest.raises(KeyError) as e:
d0.names = {"xxx": "yyy"}
assert "Cannot find column xxx in the Frame" == str(e.value)
def test_rename_bad4():
d0 = dt.Frame([[1], [2], ["hello"]], names=("a", "b", "c"))
with pytest.raises(ValueError) as e:
d0.names = ('k', )
assert ("The names list has length 1, while the Frame has 3 columns"
in str(e.value))
def test_dt_repeat2():
f0 = dt.Frame(["A", "B", "CDE"])
f1 = dt.repeat(f0, 7)
frame_integrity_check(f1)
assert f1.to_list() == [f0.to_list()[0] * 7]
def test_dt_repeat_empty_frame():
f0 = dt.Frame()
f1 = dt.repeat(f0, 5)
frame_integrity_check(f1)
assert f1.to_list() == []
def test_resize_rows_api():
f0 = dt.Frame([20])
f0.nrows = 3
f0.nrows = 5
frame_integrity_check(f0)
assert f0.to_list() == [[20, None, None, None, None]]
def test_resize_invalidates_stats():
f0 = dt.Frame([3, 1, 4, 1, 5, 9, 2, 6])
assert_equals(f0.max(), dt.Frame([9]))
f0.nrows = 3
frame_integrity_check(f0)
assert_equals(f0.max(), dt.Frame([4]))
def test_len2():
DT = dt.Frame([None, "", "mooo" * 10000], stype="str64")
RES = DT[:, f[0].len()]
assert RES.stype == dt.stype.int64
assert RES.to_list() == [[None, 0, 40000]]
def test_len_wrong_col():
DT = dt.Frame(range(34))
with pytest.raises(TypeError,
match="Function len cannot be applied to "
"a column of type int32"):
assert DT[:, f[0].len()]
def test_dt_pos(src):
DT = dt.Frame(src)
RES = DT[:, +f[0]]
stype0 = stype.int32 if DT.stype in [dt.bool8, dt.int8, dt.int16] else \
DT.stype
assert_equals(RES, dt.Frame(src, stype=stype0))
def test_len():
DT = dt.Frame(A=["", "one", "2", "three", "four", None, "six", "seventy"])
RES = DT[:, f.A.len()]
assert RES.stype == dt.stype.int64
assert RES.to_list() == [[0, 3, 1, 5, 4, None, 3, 7]]
def test_tonumpy_with_NAs_view():
# See issue #1738
X = dt.Frame(A=[5.7, 2.3, None, 4.4, 9.8, None])[1:, :]
a = X.to_numpy()
assert a.tolist() == [[2.3], [None], [4.4], [9.8], [None]]
def test_tonumpy_with_NAs4(numpy):
src = [True, False, None]
d0 = dt.Frame(src)
a0 = d0.to_numpy()
assert a0.dtype == numpy.dtype("bool")
assert a0.T.tolist() == [src]
def test_tonumpy_with_NAs3(numpy):
src = ["faa", None, "", "hooray", None]
d0 = dt.Frame(src)
a0 = d0.to_numpy()
assert a0.T.tolist() == [src]
def test_names_deduplication():
with pytest.warns(dt.exceptions.DatatableWarning):
DT = dt.Frame(names=["A", "A01", "A01"])
assert DT.names == ("A", "A01", "A2")
def test_len_unicode():
DT = dt.Frame([
"л╝л░л╣л┤л░лй", "УњЎУњѓтиеУЪњ", "ЪцЦ", "ЮћўЮћФЮћдЮћаЮћгЮћАЮћб"
])
RES = DT[:, f[0].len()]
assert_equals(RES, dt.Frame([6, 4, 1, 7], stype=dt.int64))
def test__len__():
DT = dt.Frame(A=[1, 2], B=[3, 7], D=["a", 'n'], V=[1.1, None])
assert len(DT) == 4
assert DT.__len__() == 4
with pytest.raises(TypeError):
assert DT.__len__(3)
def test_dt_neg(src):
DT = dt.Frame(src)
RES = DT[:, -f[0]]
stype0 = dt.int32 if DT.stype in [dt.bool8, dt.int8, dt.int16] else \
DT.stype
assert_equals(RES, dt.Frame([neg(x) for x in src], stype=stype0))
def test_resize_issue1527(capsys):
f0 = dt.Frame(A=[])
assert f0.nrows == 0
f0.nrows = 5
assert f0.nrows == 5
assert f0.to_list() == [[None] * 5]
def test_materialize_to_memory_bad_type():
DT = dt.Frame(range(5))
msg = r"Argument to_memory in Frame.materialize\(\) should be a boolean"
with pytest.raises(TypeError, match=msg):
DT.materialize(to_memory=0)
def test_dt_repeat():
f0 = dt.Frame(range(10))
f1 = dt.repeat(f0, 3)
frame_integrity_check(f1)
assert f1.to_list() == [list(range(10)) * 3]
def test_internal_rowindex():
d0 = dt.Frame(list(range(100)))
d1 = d0[:20, :]
assert frame_columns_virtual(d0) == [False]
assert frame_columns_virtual(d1) == [True]
def test_dt_repeat_view():
f0 = dt.Frame(A=[1, 3, 4, 5], B=[2, 6, 3, 1])
f1 = f0[::2, :]
f2 = dt.repeat(f1, 5)
frame_integrity_check(f2)
assert f2.to_dict() == {"A": [1, 4] * 5, "B": [2, 3] * 5}
def test_issue2269():
import ctypes
DT = dt.Frame(range(10000))
ptr = dt.internal.frame_column_data_r(DT, 0)
assert isinstance(ptr, ctypes.c_void_p)
def test_rename_bad1():
d0 = dt.Frame([[1], [2], ["hello"]], names=("a", "b", "c"))
with pytest.raises(TypeError):
d0.names = {"a", "b"}
def test_sizeof():
DT1 = dt.Frame(A=["foo"])
DT2 = dt.Frame(A=["foo" * 1001])
assert sys.getsizeof(DT2) - sys.getsizeof(DT1) == 3000
def test_rename_bad3():
d0 = dt.Frame([[1], [2], ["hello"]], names=("a", "b", "c"))
with pytest.raises(ValueError) as e:
d0.names = ['1', '2', '3', '5']
assert ("The names list has length 4, while the Frame has only 3 columns"
in str(e.value))
def test_dt_getitem3():
DT = dt.Frame([3])
a = DT[0]
assert a.to_list() == [[3]]
del a # should not emit an error
def test_rename_bad5():
d0 = dt.Frame([[1], [2], ["hello"]], names=("a", "b", "c"))
with pytest.raises(TypeError) as e:
d0.names = {"a": 17}
assert ("The replacement name for column a should be a string, "
"but got <class 'int'>" == str(e.value))
def test_tonumpy_with_NAs1(numpy):
src = [1, 5, None, 187, None, 103948]
d0 = dt.Frame(src)
a0 = d0.to_numpy()
assert a0.T.tolist() == [src]
