implement _local_timestamps in DatetimeArrayMixin #21721
Conversation
Codecov Report
@@ Coverage Diff @@
## master #21721 +/- ##
==========================================
+ Coverage 91.93% 91.93% +<.01%
==========================================
Files 159 159
Lines 49719 49722 +3
==========================================
+ Hits 45707 45710 +3
Misses 4012 4012
Continue to review full report at Codecov.
|
pandas/core/indexes/datetimes.py
Outdated
| reverse = np.empty(n, dtype=np.int_) | ||
| reverse.put(indexer, np.arange(n)) | ||
| return result.take(reverse) | ||
| # fall back to non-optimized implementation |
There was a problem hiding this comment.
I would remove this comment
pandas/core/arrays/datetimes.py
Outdated
| # Timezone Conversion and Localization Methods | ||
|
|
||
| def _local_timestamps(self): | ||
| values = self.asi8 |
There was a problem hiding this comment.
add a doc-string here (e.g. what are the return guarantees here)
| """ | ||
| values = self.asi8 | ||
| indexer = values.argsort() | ||
| result = conversion.tz_convert(values.take(indexer), utc, self.tz) |
There was a problem hiding this comment.
Does this only make sense if tz_convert requires its input to be sorted? I don't think it does.
There was a problem hiding this comment.
yeah not sure why that indexer is there
There was a problem hiding this comment.
In the next pass I'll see if this can be simplified.
Broken out of #21715 because it is a blocker for a handful of orthogonal follow-ons.