Skip to content

REF: dispatch DTI/TDI setops to RangeIndex #44039

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.

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 3 commits into from
Oct 16, 2021
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
75 changes: 58 additions & 17 deletions pandas/core/indexes/datetimelike.py
Original file line number Diff line number Diff line change
Expand Up @@ -28,6 +28,7 @@
Resolution,
Tick,
parsing,
to_offset,
)
from pandas.compat.numpy import function as nv
from pandas.util._decorators import (
Expand Down Expand Up @@ -61,6 +62,7 @@
NDArrayBackedExtensionIndex,
inherit_names,
)
from pandas.core.indexes.range import RangeIndex
from pandas.core.tools.timedeltas import to_timedelta

if TYPE_CHECKING:
Expand Down Expand Up @@ -433,12 +435,61 @@ def values(self) -> np.ndarray:
# --------------------------------------------------------------------
# Set Operation Methods

@cache_readonly
def _as_range_index(self) -> RangeIndex:
# Convert our i8 representations to RangeIndex
# Caller is responsible for checking isinstance(self.freq, Tick)
freq = cast(Tick, self.freq)
tick = freq.delta.value
rng = range(self[0].value, self[-1].value + tick, tick)
return RangeIndex(rng)

def _can_range_setop(self, other):
return isinstance(self.freq, Tick) and isinstance(other.freq, Tick)

def _wrap_range_setop(self, other, res_i8):
new_freq = None
if not len(res_i8):
# RangeIndex defaults to step=1, which we don't want.
new_freq = self.freq
elif isinstance(res_i8, RangeIndex):
new_freq = to_offset(Timedelta(res_i8.step))
res_i8 = res_i8

# TODO: we cannot just do
# type(self._data)(res_i8.values, dtype=self.dtype, freq=new_freq)
# because test_setops_preserve_freq fails with _validate_frequency raising.
# This raising is incorrect, as 'on_freq' is incorrect. This will
# be fixed by GH#41493
res_values = res_i8.values.view(self._data._ndarray.dtype)
result = type(self._data)._simple_new(
res_values, dtype=self.dtype, freq=new_freq
)
return self._wrap_setop_result(other, result)

def _range_intersect(self, other, sort):
# Dispatch to RangeIndex intersection logic.
left = self._as_range_index
right = other._as_range_index
res_i8 = left.intersection(right, sort=sort)
return self._wrap_range_setop(other, res_i8)

def _range_union(self, other, sort):
# Dispatch to RangeIndex union logic.
left = self._as_range_index
right = other._as_range_index
res_i8 = left.union(right, sort=sort)
return self._wrap_range_setop(other, res_i8)

def _intersection(self, other: Index, sort=False) -> Index:
"""
intersection specialized to the case with matching dtypes and both non-empty.
"""
other = cast("DatetimeTimedeltaMixin", other)

if self._can_range_setop(other):
return self._range_intersect(other, sort=sort)

if not self._can_fast_intersect(other):
result = Index._intersection(self, other, sort=sort)
# We need to invalidate the freq because Index._intersection
Expand All @@ -453,7 +504,6 @@ def _intersection(self, other: Index, sort=False) -> Index:
return self._fast_intersect(other, sort)

def _fast_intersect(self, other, sort):

# to make our life easier, "sort" the two ranges
if self[0] <= other[0]:
left, right = self, other
Expand Down Expand Up @@ -485,19 +535,9 @@ def _can_fast_intersect(self: _T, other: _T) -> bool:
# Because freq is not None, we must then be monotonic decreasing
return False

elif self.freq.is_anchored():
# this along with matching freqs ensure that we "line up",
# so intersection will preserve freq
# GH#42104
return self.freq.n == 1

elif isinstance(self.freq, Tick):
# We "line up" if and only if the difference between two of our points
# is a multiple of our freq
diff = self[0] - other[0]
remainder = diff % self.freq.delta
return remainder == Timedelta(0)

# this along with matching freqs ensure that we "line up",
# so intersection will preserve freq
# Note we are assuming away Ticks, as those go through _range_intersect
# GH#42104
return self.freq.n == 1

Expand All @@ -516,6 +556,7 @@ def _can_fast_union(self: _T, other: _T) -> bool:
return False

if len(self) == 0 or len(other) == 0:
# only reached via union_many
return True

# to make our life easier, "sort" the two ranges
Expand Down Expand Up @@ -544,10 +585,7 @@ def _fast_union(self: _TDT, other: _TDT, sort=None) -> _TDT:
loc = right.searchsorted(left_start, side="left")
right_chunk = right._values[:loc]
dates = concat_compat((left._values, right_chunk))
# With sort being False, we can't infer that result.freq == self.freq
# TODO: no tests rely on the _with_freq("infer"); needed?
result = type(self)._simple_new(dates, name=self.name)
result = result._with_freq("infer")
return result
else:
left, right = other, self
Expand All @@ -573,6 +611,9 @@ def _union(self, other, sort):
assert isinstance(other, type(self))
assert self.dtype == other.dtype

if self._can_range_setop(other):
return self._range_union(other, sort=sort)

if self._can_fast_union(other):
result = self._fast_union(other, sort=sort)
# in the case with sort=None, the _can_fast_union check ensures
Expand Down
5 changes: 4 additions & 1 deletion pandas/core/indexes/range.py
Original file line number Diff line number Diff line change
Expand Up @@ -562,8 +562,11 @@ def _intersection(self, other: Index, sort=False):

if (self.step < 0 and other.step < 0) is not (new_index.step < 0):
new_index = new_index[::-1]

if sort is None:
new_index = new_index.sort_values()
# TODO: can revert to just `if sort is None` after GH#43666
if new_index.step < 0:
new_index = new_index[::-1]

return new_index

Expand Down