bool(Dataset(False)) is True

[aaronspring]

What happened:

v=True
bool(xr.DataArray(v)) # True
bool(xr.DataArray(v).to_dataset(name='var')) # True

v=False
bool(xr.DataArray(v)) # False
# unexpected behaviour below
bool(xr.DataArray(v).to_dataset(name='var')) # True

What you expected to happen:

bool(xr.DataArray(False).to_dataset(name='var')) # False

Maybe this is intentional and I dont understand why.

xr.version = '0.16.0'

[keewis]

If I remember correctly, Dataset inherits from dict and doesn't define it's own __bool__, so it has the behavior of dict.__bool__:

In [2]: bool(xr.Dataset())
Out[2]: False

In [3]: bool(xr.Dataset({"x": False}))
Out[3]: True

In [4]: bool({})
Out[4]: False

In [5]: bool({"x": False})
Out[5]: True

[aaronspring]

thanks for the explanation. but from a naive user perspective,

bool(xr.DataArray(False).to_dataset(name='var')) # False
bool(xr.DataArray(True).to_dataset(name='var')) # True
bool(xr.DataArray(False)) # False
bool(xr.DataArray(True)) # True

would be nice.

[keewis]

I'm not sure if that's something we want to do: what happens if the dataset has more than one variable?

I guess what we're talking about here is making bool(ds) a alias of all(ds.values() or [False]) (the ... or [False] is for making Dataset({}) return False instead of True).

Thoughts, @pydata/xarray?

[aaronspring]

currently

if (xrobject > 10).any():
   raise ValueError('blabla')

depends on whether xrobject is xr.DataArray or xr.Dataset.

[TomNicholas]

This question seems to me to be conceptually similar to the question I described in #3315: "Should we treat a named DataArray and a single-variable Dataset as if they are the same?"

If the answer is yes, then we should definitely change the behaviour of bool(Dataset()) as Aaron suggests. If instead we decide "no, Datasets are dictionaries of DataArrays, and a single-element dictionary is not the same thing as the named object it contains", then I'm not sure.

Intuitively I personally think it should be the former, but I would first like to see how many tests break if we make that kind of change.

[max-sixty]

We mirror numpy for DataArray and, per @keewis , Mapping for Dataset. I think we should have a fairly high bar for deviating from those.

For those who think bool(xr.Dataset(dict(x=False))) should evaluate to False, what do you think dict(x=False) should evaluate to?

IIUC, this is somewhat of an edge case — it's only possible to coerce DataArray to bool when there's exactly one element. (Not that we shouldn't resolve)

This question seems to me to be conceptually similar to the question I described in #3315: "Should we treat a named DataArray and a single-variable Dataset as if they are the same?"

👍

[dcherian]

We mirror numpy for DataArray and, per @keewis , Mapping for Dataset. I think we should have a fairly high bar for deviating from those.

👍

[aaronspring]

For those who think bool(xr.Dataset(dict(x=False))) should evaluate to False, what do you think dict(x=False) should evaluate to?

good point

IIUC, this is somewhat of an edge case — it's only possible to coerce DataArray to bool when there's exactly one element. (Not that we shouldn't resolve)

I now understand the underlying problem better with your comments. My expectations were too naive about it. I cannot evaluate the consequences of changing this. I close this issue and hopefully users having this problem in the future will find this issue.

[max-sixty]

Thanks for raising and engaging @aaronspring !