quantile: rename interpolation arg to method #6108
Conversation
|
|
xarray/core/dataarray.py
Outdated
| - nearest: ``i`` or ``j``, whichever is nearest. | ||
| - midpoint: ``(i + j) / 2``. | ||
| use when the desired quantile lies between two data points. | ||
| See numpy.quantile for available methods. |
There was a problem hiding this comment.
Is this ok or should we list them explicitly? (The available methods depend on the numpy version which makes this a bit difficult).
There was a problem hiding this comment.
I would lean toward listing the current available methods but if not, we should make sure numpy.quantile get's picked up as a link by intersphinx.
|
+1 on making this change. However, just to note, Pandas seems to still have the the |
xarray/core/dataarray.py
Outdated
| interpolation: str = "linear", | ||
| q: ArrayLike, | ||
| dim: str | Sequence[Hashable] | None = None, | ||
| method: str = "linear", |
There was a problem hiding this comment.
Can use Literal["linear", etc] here instead of str.
There was a problem hiding this comment.
I feel that's a bit over the top - can I get away without it?
There was a problem hiding this comment.
Agree it would be cool, definitely fine without out too! :)
|
The rtd failure looks unrelated. |
| 8. "median_unbiased" (*) | ||
| 9. "normal_unbiased" (*) | ||
|
|
||
| The first three methods are discontiuous. The following discontinuous |
numpy/numpy#20327 introduces some changes to
np.quantile(and related) for the upcoming numpy release (v1.22.0). It renames theinterpolationkeyword tomethodand offers some new interpolation methods. This PR does two thingsinterpolationkeyword tomethodin xarray - this change is not strictly necessary but I thought better to be consistent with numpypre-commit run --all-fileswhats-new.rst(Side note in
dask.array.percentilethemethodkeyword is used differently from theinterpolationkeyword (docs). However, xarray does not use the dask function.)TODO: need to import
ArrayLikefrom npcompat.