Skip to content

REF: share to_native_types with ArrayManager #40490

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Mar 19, 2021
Merged

REF: share to_native_types with ArrayManager #40490

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2cd4b39
REF: share to_native_types with ArrayManager
jbrockmendel Mar 17, 2021
8e633fa
REF: share fill_value
jbrockmendel Mar 17, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion pandas/core/internals/array_manager.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -88,6 +88,7 @@
from pandas.core.internals.blocks import (
ensure_block_shape,
new_block,
to_native_types,
)

if TYPE_CHECKING:
Expand Down Expand Up @@ -634,7 +635,7 @@ def replace_list(
)

def to_native_types(self, **kwargs):
return self.apply_with_block("to_native_types", **kwargs)
return self.apply(to_native_types, **kwargs)

@property
def is_mixed_type(self) -> bool:
Expand Down
158 changes: 79 additions & 79 deletions pandas/core/internals/blocks.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -97,6 +97,7 @@
FloatingArray,
IntegerArray,
PandasArray,
TimedeltaArray,
)
from pandas.core.base import PandasObject
import pandas.core.common as com
Expand Down Expand Up @@ -260,9 +261,11 @@ def get_block_values_for_json(self) -> np.ndarray:
# TODO(EA2D): reshape will be unnecessary with 2D EAs
return np.asarray(self.values).reshape(self.shape)

@final
@property
def fill_value(self):
return np.nan
# Used in reindex_indexer
return na_value_for_dtype(self.dtype, compat=False)

@property
def mgr_locs(self) -> BlockPlacement:
Expand Down Expand Up @@ -652,24 +655,11 @@ def should_store(self, value: ArrayLike) -> bool:
"""
return is_dtype_equal(value.dtype, self.dtype)

@final
def to_native_types(self, na_rep="nan", quoting=None, **kwargs):
""" convert to our native types format """
values = self.values

mask = isna(values)
itemsize = writers.word_len(na_rep)

if not self.is_object and not quoting and itemsize:
values = values.astype(str)
if values.dtype.itemsize / np.dtype("U1").itemsize < itemsize:
# enlarge for the na_rep
values = values.astype(f"<U{itemsize}")
else:
values = np.array(values, dtype="object")

values[mask] = na_rep
values = values.astype(object, copy=False)
return self.make_block(values)
result = to_native_types(self.values, na_rep=na_rep, quoting=quoting, **kwargs)
return self.make_block(result)

# block actions #
@final
Expand Down Expand Up @@ -1498,11 +1488,6 @@ def _holder(self):
# For extension blocks, the holder is values-dependent.
return type(self.values)

@property
def fill_value(self):
# Used in reindex_indexer
return self.values.dtype.na_value

@property
def _can_hold_na(self):
# The default ExtensionArray._can_hold_na is True
Expand Down Expand Up @@ -1565,15 +1550,6 @@ def get_values(self, dtype: Optional[DtypeObj] = None) -> np.ndarray:
def array_values(self) -> ExtensionArray:
return self.values

def to_native_types(self, na_rep="nan", quoting=None, **kwargs):
"""override to use ExtensionArray astype for the conversion"""
values = self.values
mask = isna(values)

new_values = np.asarray(values.astype(object))
new_values[mask] = na_rep
return self.make_block(new_values)

def take_nd(
self,
indexer,
Expand Down Expand Up @@ -1808,41 +1784,6 @@ def is_bool(self):
class FloatBlock(NumericBlock):
__slots__ = ()

def to_native_types(
self, na_rep="", float_format=None, decimal=".", quoting=None, **kwargs
):
""" convert to our native types format """
values = self.values

# see gh-13418: no special formatting is desired at the
# output (important for appropriate 'quoting' behaviour),
# so do not pass it through the FloatArrayFormatter
if float_format is None and decimal == ".":
mask = isna(values)

if not quoting:
values = values.astype(str)
else:
values = np.array(values, dtype="object")

values[mask] = na_rep
values = values.astype(object, copy=False)
return self.make_block(values)

from pandas.io.formats.format import FloatArrayFormatter

formatter = FloatArrayFormatter(
values,
na_rep=na_rep,
float_format=float_format,
decimal=decimal,
quoting=quoting,
fixed_width=False,
)
res = formatter.get_result_as_array()
res = res.astype(object, copy=False)
return self.make_block(res)


class NDArrayBackedExtensionBlock(HybridMixin, Block):
"""
Expand Down Expand Up @@ -1962,18 +1903,6 @@ def array_values(self):
def _holder(self):
return type(self.array_values())

@property
def fill_value(self):
return na_value_for_dtype(self.dtype)

def to_native_types(self, na_rep="NaT", **kwargs):
""" convert to our native types format """
arr = self.array_values()

result = arr._format_native_types(na_rep=na_rep, **kwargs)
result = result.astype(object, copy=False)
return self.make_block(result)


class DatetimeBlock(DatetimeLikeBlockMixin):
__slots__ = ()
Expand All @@ -1999,7 +1928,6 @@ class DatetimeTZBlock(ExtensionBlock, DatetimeBlock):

internal_values = Block.internal_values
_can_hold_element = DatetimeBlock._can_hold_element
to_native_types = DatetimeBlock.to_native_types
diff = DatetimeBlock.diff
where = DatetimeBlock.where
putmask = DatetimeLikeBlockMixin.putmask
Expand Down Expand Up @@ -2316,3 +2244,75 @@ def ensure_block_shape(values: ArrayLike, ndim: int = 1) -> ArrayLike:
# We can't, and don't need to, reshape.
values = np.asarray(values).reshape(1, -1)
return values


def to_native_types(
values: ArrayLike,
*,
na_rep="nan",
quoting=None,
float_format=None,
decimal=".",
**kwargs,
) -> np.ndarray:
""" convert to our native types format """
values = ensure_wrapped_if_datetimelike(values)

if isinstance(values, (DatetimeArray, TimedeltaArray)):
result = values._format_native_types(na_rep=na_rep, **kwargs)
result = result.astype(object, copy=False)
return result

elif isinstance(values, ExtensionArray):
mask = isna(values)

new_values = np.asarray(values.astype(object))
new_values[mask] = na_rep
return new_values

elif values.dtype.kind == "f":
# see GH#13418: no special formatting is desired at the
# output (important for appropriate 'quoting' behaviour),
# so do not pass it through the FloatArrayFormatter
if float_format is None and decimal == ".":
mask = isna(values)

if not quoting:
values = values.astype(str)
else:
values = np.array(values, dtype="object")

values[mask] = na_rep
values = values.astype(object, copy=False)
return values

from pandas.io.formats.format import FloatArrayFormatter

formatter = FloatArrayFormatter(
values,
na_rep=na_rep,
float_format=float_format,
decimal=decimal,
quoting=quoting,
fixed_width=False,
)
res = formatter.get_result_as_array()
res = res.astype(object, copy=False)
return res

else:

mask = isna(values)
itemsize = writers.word_len(na_rep)

if values.dtype != _dtype_obj and not quoting and itemsize:
values = values.astype(str)
if values.dtype.itemsize / np.dtype("U1").itemsize < itemsize:
# enlarge for the na_rep
values = values.astype(f"<U{itemsize}")
else:
values = np.array(values, dtype="object")

values[mask] = na_rep
values = values.astype(object, copy=False)
return values