Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Pint support for DataArray (pydata#3643)
* remove xfail marks from median and cumprod * remove all xfails not related to indexes or external packages * switch away from using assert_equal_with_units * use assert_allclose in a few cases instead * don't use a kwarg for searchsorted normally, this should work, but the documentation mismatches the implementation of searchsorted and names the keys as `keys` instead of `v` * move the tests for item into their own test function * move the searchsorted tests into their own test function * remove a wrapping pytest.param * treat objects implementing __array_function__ the same as ndarray * mark numpy.median as xfailing * remove the xfail marks for the all and any tests * use assert_units_equal to check the resulting units * don't attempt to use interpolate_na with int dtype arrays * update the xfail reason for DataArray.interpolate_na * xfail the compatible units bivariate_ufunc test and don't use 0 * combine and expand the reindex and interp tests * combine and expand the reindex_like and interp_like tests * xfail the quantile tests if pint is not recent enough * xfail the rolling tests * don't xfail combine_first it currently does not test indexing, so probably will need a new test for that. * use numpy's assert_allclose * don't add dimension coordinates if they're not necessary * add the PR to the list of related PRs * move the whats-new.rst entry to 0.16.0 * check for __array_ufunc__ to decide if the type is supported * xfail the bivariate ufunc tests * remove the check for __array_ufunc__ * skip the DataArray.identical tests * use pytest.param
- Loading branch information