pydata / xarray / 5728

96%

push

travis-ci

Refactor nanops (#2236)

* Inhouse nanops

* Cleanup nanops

* remove NAT_TYPES

* flake8.

* another flake8

* recover nat types

* remove keep_dims option from nanops (to make them compatible with numpy==1.11).

* Test aggregation over multiple dimensions

* Remove print.

* Docs. More cleanup.

* flake8

* Bug fix. Better test coverage.

* using isnull, where_method. Remove unnecessary conditional branching.

* More refactoring based on the comments

* remove dtype from nanmedian

* Fix for nanmedian

* Add tests for dataset

* Add tests with resample.

* lint

* updated whatsnew

* Revise from comments.

* Use .any and .all method instead of np.any / np.all

* Avoid using numpy methods

* Avoid casting to int for dask array

* Update whatsnew

273 of 273 new or added lines in 10 files covered. (100.0%)

24928 of 25716 relevant lines covered (96.94%)

8.17 hits per line