xarray.Dataset.var - xarray.DataArray.var - does it have ddof=1 parameter?

[chiaral]

It is not clear from the description whether ddof = 1 is available and/or if it is set to 0.
(https://docs.scipy.org/doc/numpy-1.6.0/reference/generated/numpy.var.html)

for large samples, 1 or 0 don't make a lot of difference, but it would be good to know whether it uses N-1 or N.

[shoyer]

Good question. Setting ddof should work. It's passed on to nanvar from NumPy or bottleneck, both of which default to ddof=0.

[stale]

In order to maintain a list of currently relevant issues, we mark issues as stale after a period of inactivity
If this issue remains relevant, please comment here; otherwise it will be marked as closed automatically

[sjvrijn]

In core/nanops.py there are some explicit defaults of ddof=0 within xarray, but I'm not sure if those are always used or if there are also cases where var (or std) are directly passed on to numpy/bottleneck/dask.

I'm considering two different options to clarify this:

1. Add a docstring section on the ddof parameter specifying it uses ddof=0 as default for the reduction methods that use it, i.e. var and std. Possibly just copied from numpy's var page.
2. Refer to numpy's documentation page in the docstring of all reduction methods for further reference.

Both would require some logic in core/ops.py: either to check for which reduce methods need a ddof paragraph, or to create the proper url (which has to adjust min and max to np.amin and np.amax respectively)

Is there any clear preference from anyone about this?

[max-sixty]

Thanks for finding that @sjvrijn

I don't have a view on what we should use, so I would vote to defer to numpy (and pandas, which also seems to use 1), referencing that documentation to the extent xarray isn't changing anything.

But others probably have a stronger view on which ddof we should use?