This data as json
| sha | message | author_date | committer_date | raw_author | raw_committer | repo | author | committer |
|---|---|---|---|---|---|---|---|---|
| ecc442054f05c5b500f3b58d3c9dacce75a9de36 | keep_attrs for Dataset.resample and DataArray.resample (#829) * initial addition of keep_attrs argument to resample, and tests resample now has a keep_attrs argument that is passed on to reduction operations. However, it is not being applied in all cases. Some tests have been added that do not pass. * moved keep_attrs kwarg from func to apply for DataArray The funcs passed in to DataArrayGroupBy.apply did not respect the keep_attrs argument, and seem to keep attributes by default. Now DataArrayGroupBy.apply does the work of clearing result.attrs if keep_attrs is False, and does nothing to the attributes otherwise. This still needs to be reflected in the docstring. * DataArrayGroupBy concat no longer sometimes copies metadata DataArrayGroupBy._concat_shortcut copied metadata from itself to the applied arrays, when it should not. Metadata copying should be, and is, handled elsewhere. With this fix, keep_attrs for resample methods is implemented and passes the written tests. No other tests are broken. * default for resample is now keep_attrs=False, tests for Dataset.resample Tests have been added to make sure that keep_attrs applies to Dataset attributes as well as the attributes of variables in that Dataset. resample now has keep_attrs=False as its default regardless of the operation being completed (previously it was True for 'first' and 'last') * cleaned up tests for keep_attrs option of resample * added "what's new" notes for keep_attrs=False resampling option | 2016-04-19T16:02:32Z | 2016-04-19T16:02:32Z | 1460887cf8ffb300273c62012a1a458ec3bbda91 | f10b21bed2846b879806f87039b77245b18e7671 | 13221727 | 12307589 | 1217238 |