use .values in index difference

Author: max-sixty

doc/source/whatsnew/v0.17.1.txt

@@ -136,6 +136,7 @@ Performance Improvements
 ~~~~~~~~~~~~~~~~~~~~~~~~
 
 - Checking monotonic-ness before sorting on an index (:issue:`11080`)
+- Performance improvements in ``Index.difference``, particularly for ``PeriodIndex`` (:issue:`11278`)
 - ``Series.dropna`` performance improvement when its dtype can't contain ``NaN`` (:issue:`11159`)
 - Release the GIL on most datetime field operations (e.g. ``DatetimeIndex.year``, ``Series.dt.year``), normalization, and conversion to and from ``Period``, ``DatetimeIndex.to_period`` and ``PeriodIndex.to_timestamp`` (:issue:`11263`)
 - Release the GIL on some rolling algos: ``rolling_median``, ``rolling_mean``, ``rolling_max``, ``rolling_min``, ``rolling_var``, ``rolling_kurt``, ``rolling_skew`` (:issue:`11450`)

pandas/core/index.py

@@ -1711,12 +1711,12 @@ def difference(self, other):
  self._assert_can_do_setop(other)
 
  if self.equals(other):
- return Index([], name=self.name)
+ return self._shallow_copy([])
 
  other, result_name = self._convert_can_do_setop(other)
 
- theDiff = sorted(set(self) - set(other))
- return Index(theDiff, name=result_name)
+ diff = sorted(set(self.values) - set(other.values))
+ return self._shallow_copy(diff, name=result_name)
 
  diff = deprecate('diff', difference)