def test_scalar_add_invalid(dtype_l, dtype_r):
test_value = 1
lval_gpu = cudf.Scalar(test_value, dtype=dtype_l)
rval_gpu = cudf.Scalar(test_value, dtype=dtype_r)
with pytest.raises(TypeError):
lval_gpu + rval_gpu
def test_boolean_scalar_binop(op):
psr = pd.Series(np.random.choice([True, False], 10))
gsr = cudf.from_pandas(psr)
utils.assert_eq(op(psr, True), op(gsr, True))
utils.assert_eq(op(psr, False), op(gsr, False))
# cuDF scalar
utils.assert_eq(op(psr, True), op(gsr, cudf.Scalar(True)))
utils.assert_eq(op(psr, False), op(gsr, cudf.Scalar(False)))
def test_scalar_dtype_and_validity_decimal(slr, dtype, expect):
if expect is pa.lib.ArrowInvalid:
with pytest.raises(expect):
cudf.Scalar(slr, dtype=dtype)
return
else:
result = cudf.Scalar(slr, dtype=dtype)
assert result.dtype == dtype
assert result.is_valid
def test_null_scalar(dtype):
if isinstance(dtype, cudf.Decimal64Dtype):
with pytest.raises(NotImplementedError):
s = cudf.Scalar(None, dtype=dtype)
return
s = cudf.Scalar(None, dtype=dtype)
assert s.value is cudf.NA
assert s.dtype == np.dtype(dtype)
assert s.is_valid() is False
def test_scalar_dtype_and_validity(dtype):
if isinstance(dtype, cudf.Decimal64Dtype):
with pytest.raises(NotImplementedError):
s = cudf.Scalar(None, dtype=dtype)
return
s = cudf.Scalar(1, dtype=dtype)
assert s.dtype == np.dtype(dtype)
assert s.is_valid() is True
def __setitem__(self, key, value):
if isinstance(value, list):
value = cudf.Scalar(value)
if isinstance(value, cudf.Scalar):
if value.dtype != self.dtype:
raise TypeError("list nesting level mismatch")
elif value is cudf.NA:
value = cudf.Scalar(value, dtype=self.dtype)
else:
raise ValueError(f"Can not set {value} into ListColumn")
super().__setitem__(key, value)
def test_construct_from_scalar(value):
value = cudf.utils.utils.to_cudf_compatible_scalar(value)
x = cudf.Scalar(value, value.dtype)
y = cudf.Scalar(x)
assert x.value == y.value or np.isnan(x.value) and np.isnan(y.value)
# check that this works:
y.device_value
x._is_host_value_current == y._is_host_value_current
x._is_device_value_current == y._is_device_value_current
def test_scalar_difference(dtype_l, dtype_r):
test_value = 1
lval_host = dtype_scalar(test_value, dtype=dtype_l)
rval_host = dtype_scalar(test_value, dtype=dtype_r)
lval_gpu = cudf.Scalar(test_value, dtype=dtype_l)
rval_gpu = cudf.Scalar(test_value, dtype=dtype_r)
expect = lval_host - rval_host
got = lval_gpu - rval_gpu
assert expect == got.value
assert expect.dtype == got.dtype
def test_scalar_floordiv(dtype_l, dtype_r):
test_value = 1
lval_host = dtype_scalar(test_value, dtype=dtype_l)
rval_host = dtype_scalar(test_value, dtype=dtype_r)
lval_gpu = cudf.Scalar(test_value, dtype=dtype_l)
rval_gpu = cudf.Scalar(test_value, dtype=dtype_r)
expect = lval_host // rval_host
got = lval_gpu // rval_gpu
assert expect == got.value
assert expect.dtype == got.dtype
def test_scalar_power(dtype_l, dtype_r):
test_value = 1
lval_host = dtype_scalar(test_value, dtype=dtype_l)
rval_host = dtype_scalar(test_value, dtype=dtype_r)
lval_gpu = cudf.Scalar(test_value, dtype=dtype_l)
rval_gpu = cudf.Scalar(test_value, dtype=dtype_r)
expect = lval_host**rval_host
got = lval_gpu**rval_gpu
assert expect == got.value
assert expect.dtype == got.dtype
def _check_and_cast_columns_with_other(
source_col: ColumnBase,
other: Union[ScalarLike, ColumnBase],
inplace: bool,
) -> Tuple[ColumnBase, Union[ScalarLike, ColumnBase]]:
"""
Returns type-casted column `source_col` & scalar `other_scalar`
based on `inplace` parameter.
"""
if cudf.utils.dtypes.is_categorical_dtype(source_col.dtype):
return source_col, other
if cudf.utils.dtypes.is_scalar(other):
device_obj = _normalize_scalars(source_col, other)
else:
device_obj = other
if other is None:
return source_col, device_obj
elif cudf.utils.dtypes.is_mixed_with_object_dtype(device_obj, source_col):
raise TypeError("cudf does not support mixed types, please type-cast "
"the column of dataframe/series and other "
"to same dtypes.")
if inplace:
if not cudf.utils.dtypes._can_cast(device_obj.dtype, source_col.dtype):
warnings.warn(
f"Type-casting from {device_obj.dtype} "
f"to {source_col.dtype}, there could be potential data loss")
return source_col, device_obj.astype(source_col.dtype)
else:
if (cudf.utils.dtypes.is_scalar(other)
and cudf.utils.dtypes.is_numerical_dtype(source_col.dtype)
and cudf.utils.dtypes._can_cast(other, source_col.dtype)):
common_dtype = source_col.dtype
return (
source_col.astype(common_dtype),
cudf.Scalar(other, dtype=common_dtype),
)
else:
common_dtype = cudf.utils.dtypes.find_common_type([
source_col.dtype,
np.min_scalar_type(other)
if cudf.utils.dtypes.is_scalar(other) else other.dtype,
])
if cudf.utils.dtypes.is_scalar(device_obj):
device_obj = cudf.Scalar(other, dtype=common_dtype)
else:
device_obj = device_obj.astype(common_dtype)
return source_col.astype(common_dtype), device_obj
def test_scalar_add(dtype_l, dtype_r):
test_value = 1
lval_host = dtype_scalar(test_value, dtype=dtype_l)
rval_host = dtype_scalar(test_value, dtype=dtype_r)
lval_gpu = cudf.Scalar(test_value, dtype=dtype_l)
rval_gpu = cudf.Scalar(test_value, dtype=dtype_r)
# expect = np.add(lval_host, rval_host)
expect = lval_host + rval_host
got = lval_gpu + rval_gpu
assert expect == got.value
if not dtype_l == dtype_r == "str":
assert expect.dtype == got.dtype
def fillna(
self, value: Any = None, method: str = None, dtype: Dtype = None
):
"""Fill null values with ``value``.
Returns a copy with null filled.
"""
if isinstance(value, (int, Decimal)):
value = cudf.Scalar(value, dtype=self.dtype)
elif (
isinstance(value, Decimal64Column)
or isinstance(value, cudf.core.column.NumericalColumn)
and is_integer_dtype(value.dtype)
):
value = value.astype(self.dtype)
else:
raise TypeError(
"Decimal columns only support using fillna with decimal and "
"integer values"
)
result = libcudf.replace.replace_nulls(
input_col=self, replacement=value, method=method, dtype=dtype
)
return result._with_type_metadata(self.dtype)
def _binary_op_truediv(
self, rhs: BinaryOperand
) -> Tuple["column.ColumnBase", BinaryOperand, DtypeObj]:
lhs = self # type: column.ColumnBase
if pd.api.types.is_timedelta64_dtype(rhs.dtype):
common_dtype = determine_out_dtype(self.dtype, rhs.dtype)
lhs = lhs.astype(common_dtype).astype("float64")
if isinstance(rhs, cudf.Scalar):
if rhs.is_valid():
rhs = rhs.value.astype(common_dtype).astype("float64")
else:
rhs = cudf.Scalar(None, "float64")
else:
rhs = rhs.astype(common_dtype).astype("float64")
out_dtype = np.dtype("float64")
elif rhs.dtype.kind in ("f", "i", "u"):
out_dtype = self.dtype
else:
raise TypeError(
f"Division of {self.dtype} with {rhs.dtype} "
f"cannot be performed."
)
return lhs, rhs, out_dtype
def test_list_scalar_host_construction_null(elem_type, nesting_level):
dtype = cudf.ListDtype(elem_type)
for level in range(nesting_level - 1):
dtype = cudf.ListDtype(dtype)
slr = cudf.Scalar(None, dtype=dtype)
assert slr.value is cudf.NA
def test_scalar_host_initialization(value):
s = cudf.Scalar(value)
np.testing.assert_equal(s.value, value)
assert s.is_valid() is True
assert s._is_host_value_current
assert not s._is_device_value_current
def test_scalar_logical():
T = cudf.Scalar(True)
F = cudf.Scalar(False)
assert T
assert not F
assert T and T
assert not (T and F)
assert not (F and T)
assert not (F and F)
assert T or T
assert T or F
assert F or T
assert not (F or F)
def fillna(self, fill_value=None, method=None):
"""
Fill null values with *fill_value*
"""
if method is not None:
return super().fillna(fill_value, method)
if (isinstance(fill_value, cudf.Scalar)
and fill_value.dtype == self.dtype):
return super().fillna(fill_value, method)
if np.isscalar(fill_value):
# castsafely to the same dtype as self
fill_value_casted = self.dtype.type(fill_value)
if not np.isnan(fill_value) and (fill_value_casted != fill_value):
raise TypeError(
f"Cannot safely cast non-equivalent "
f"{type(fill_value).__name__} to {self.dtype.name}")
fill_value = cudf.Scalar(fill_value_casted)
else:
fill_value = column.as_column(fill_value, nan_as_null=False)
# cast safely to the same dtype as self
if is_integer_dtype(self.dtype):
fill_value = _safe_cast_to_int(fill_value, self.dtype)
else:
fill_value = fill_value.astype(self.dtype)
return super().fillna(fill_value, method)
def contains(self, search_key: ScalarLike) -> ParentType:
"""
Returns boolean values indicating whether the specified scalar
is an element of each row.
Parameters
----------
search_key : scalar
element being searched for in each row of the list column
Returns
-------
Series or Index
Examples
--------
>>> s = cudf.Series([[1, 2, 3], [3, 4, 5], [4, 5, 6]])
>>> s.list.contains(4)
Series([False, True, True])
dtype: bool
"""
search_key = cudf.Scalar(search_key)
try:
res = self._return_or_inplace(
contains_scalar(self._column, search_key))
except RuntimeError as e:
if ("Type/Scale of search key does not"
"match list column element type" in str(e)):
raise TypeError("Type/Scale of search key does not"
"match list column element type") from e
raise
else:
return res
def test_timedelta_index_ops_with_cudf_scalars(data, cpu_scalar, dtype, op):
gtdi = cudf.Index(data=data, dtype=dtype)
ptdi = gtdi.to_pandas()
gpu_scalar = cudf.Scalar(cpu_scalar)
if op == "add":
expected = ptdi + cpu_scalar
actual = gtdi + gpu_scalar
elif op == "sub":
expected = ptdi - cpu_scalar
actual = gtdi - gpu_scalar
elif op == "truediv":
expected = ptdi / cpu_scalar
actual = gtdi / gpu_scalar
elif op == "floordiv":
expected = ptdi // cpu_scalar
actual = gtdi // gpu_scalar
assert_eq(expected, actual)
if op == "add":
expected = cpu_scalar + ptdi
actual = gpu_scalar + gtdi
elif op == "sub":
expected = cpu_scalar - ptdi
actual = gpu_scalar - gtdi
elif op == "truediv":
expected = cpu_scalar / ptdi
actual = gpu_scalar / gtdi
elif op == "floordiv":
expected = cpu_scalar // ptdi
actual = gpu_scalar // gtdi
assert_eq(expected, actual)
def test_scalar_no_negative_bools():
x = cudf.Scalar(True)
with pytest.raises(
TypeError,
match=re.escape("Boolean scalars in cuDF do not "
"support negation, use logical not"),
):
-x
def seconds(self) -> "cudf.core.column.NumericalColumn":
"""
Number of seconds (>= 0 and less than 1 day).
Returns
-------
NumericalColumn
"""
# This property must return the number of seconds (>= 0 and
# less than 1 day) for each element, hence first performing
# mod operation to remove the number of days and then performing
# division operation to extract the number of seconds.
return (self % cudf.Scalar(
np.timedelta64(
_numpy_to_pandas_conversion["D"], "ns"))) // cudf.Scalar(
np.timedelta64(_numpy_to_pandas_conversion["s"], "ns"))
def __setitem__(self, key, value):
if isinstance(value, dict):
# filling in fields not in dict
for field in self.dtype.fields:
value[field] = value.get(field, cudf.NA)
value = cudf.Scalar(value, self.dtype)
super().__setitem__(key, value)
def normalize_binop_value(self, other):
if is_scalar(other) and isinstance(other, (int, np.int, Decimal)):
return cudf.Scalar(Decimal(other))
elif isinstance(other, cudf.Scalar) and isinstance(
other.dtype, cudf.core.dtypes.DecimalDtype):
return other
else:
raise TypeError(f"cannot normalize {type(other)}")
def test_datetime_series_ops_with_cudf_scalars(data, scalar, dtype, op):
gsr = cudf.Series(data=data, dtype=dtype)
psr = gsr.to_pandas()
expect = op(psr, scalar)
got = op(gsr, cudf.Scalar(scalar))
assert_eq(expect, got)
def nanoseconds(self) -> "cudf.core.column.NumericalColumn":
"""
Return the number of nanoseconds (n), where 0 <= n < 1 microsecond.
Returns
-------
NumericalColumn
"""
# This property must return the number of nanoseconds (>= 0 and
# less than 1 microsecond) for each element, hence first performing
# mod operation to remove the number of microseconds and then
# performing division operation to extract the number
# of nanoseconds.
return (self % cudf.Scalar(
np.timedelta64(
_numpy_to_pandas_conversion["us"], "ns"))) // cudf.Scalar(
np.timedelta64(_numpy_to_pandas_conversion["ns"], "ns"))
def cast_to_appropriate_cudf_type(val):
# TODO Handle scalar
if val.ndim == 0:
return cudf.Scalar(val).value
# 1D array
elif (val.ndim == 1) or (val.ndim == 2 and val.shape[1] == 1):
return cudf.Series(val)
else:
return NotImplemented
def test_series_iloc_defer_cudf_scalar():
ps = pd.Series([1, 2, 3], index=pd.Index(["a", "b", "c"]))
gs = cudf.from_pandas(ps)
for t in index_dtypes:
arg = cudf.Scalar(1, dtype=t)
got = gs[arg]
expect = 2
assert_eq(expect, got)
def test_null_scalar(dtype):
s = cudf.Scalar(None, dtype=dtype)
assert s.value is cudf.NA
assert s.dtype == (
cudf.dtype(dtype)
if not isinstance(dtype, cudf.core.dtypes.DecimalDtype)
else dtype
)
assert s.is_valid() is False
def days(self) -> "cudf.core.column.NumericalColumn":
"""
Number of days for each element.
Returns
-------
NumericalColumn
"""
return self // cudf.Scalar(
np.timedelta64(_numpy_to_pandas_conversion["D"], "ns"))