What is your issue?

Any ideas what could be causing these? I can't immediately reproduce locally.

https://github.com/pydata/xarray/actions/runs/7105414268/job/19342564583

``` Error: TestDataArray.test_computation_objects[int64-method_groupby_bins-data]

AssertionError: Left and right DataArray objects are not close

Differing values: L <Quantity([[ nan nan 1. 1. ] [2. 2. 3. 3. ] [4. 4. 5. 5. ] [6. 6. 7. 7. ] [8. 8. 9. 9.333333]], 'meter')> R <Quantity([[0. 0. 1. 1. ] [2. 2. 3. 3. ] [4. 4. 5. 5. ] [6. 6. 7. 7. ] [8. 8. 9. 9.333333]], 'meter')> ```

I was doing some profiling on isel, and see there are some properties that (I think) never change, but are called frequently. Should we cache these on their object?

Pandas uses cache_readonly for these cases.

Here's a case: we call LazilyOuterIndexedArray.shape frequently when doing a simple indexing operation. Each call takes ~150µs. An attribute lookup on a python object takes ~50ns (i.e. 3000x faster). IIUC the result on that property should never change.

I don't think this is the solution to performance issues, and there's some additional complexity. Could they be easy & small wins, though?

I think with np.flip and bn.push, this should be simple. They're both fairly new and so would require version checks / upgrading the minimums.

One small issue, I wonder if anyone has come across this: bottleneck returns the numpy array rather than the DataArray - is that because it's not operating with the correct numpy interface?

Forward fill: array.values = bn.push(array.values, axis=array.get_axis_num(axis_name))

Backfill: ``` axis = array.get_axis_num(axis_name)

reverse for bfill

array = np.flip(array, axis=axis)

fill

array.values = bn.push(array.values, axis=axis)

reverse back to original

result = np.flip(scaling, axis=date_axis) ```

This looks like it's available with dir / tab complete, but actually isn't:

``` python In [13]: xr.DataArray(np.random.rand(10,3)).to_dataset('dim_1').rolling

AttributeError Traceback (most recent call last) <ipython-input-13-438d3638a0d0> in <module>() ----> 1 xr.DataArray(np.random.rand(10,3)).to_dataset('dim_1').rolling

/Library/Frameworks/Python.framework/Versions/3.5/lib/python3.5/site-packages/xarray/core/common.py in getattr(self, name) 135 return source[name] 136 raise AttributeError("%r object has no attribute %r" % --> 137 (type(self).name, name)) 138 139 def setattr(self, name, value):

AttributeError: 'Dataset' object has no attribute 'rolling' ```

I think this could be easy to implement as an .apply operation? (indeed, that could be a reasonable path for a whole host of operations - i.e. try and apply them to each array in the ds?)

Also, as a very narrow point, I'm not sure why .rolling_cls is public? Probably should be private?

Finally, the Rolling implementation is pretty sweet. I've been getting my hands dirty in the pandas one recently, and that we can have something as well featured as that with so few lines of code 👍

This is very peculiar & specific, but also fairly impactful for us.

If you - Create a Dataset with a coord that is a PeriodIndex - Transpose that coord - Add a variable to the Dataset that needs to be reindexed

...then the type of the index changes from object to int64. This then causes other arrays added to that dataset to show up as NaNs throughout.

Here's an example. Note the dtype('O')) at the end of each output.

``` python In [61]: series = pd.Series(np.random.rand(10),index=pd.period_range(start='2000', periods=10,name='date')) ​ ds = xray.Dataset({'number 1':series}) ds['number 2'] = ds['number 1'] ds, ds.date.dtype Out[61]: (<xray.Dataset> Dimensions: (date: 10) Coordinates: * date (date) object 10957 10958 10959 10960 10961 10962 10963 10964 ... Data variables: number 1 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 0.6723 ... number 2 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 0.6723 ..., dtype('O')) In [62]:

ds, ds.date.dtype ds=ds.transpose('date') ds, ds.date.dtype Out[62]: (<xray.Dataset> Dimensions: (date: 10) Coordinates: * date (date) object 10957 10958 10959 10960 10961 10962 10963 10964 ... Data variables: number 1 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 0.6723 ... number 2 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 0.6723 ..., dtype('O')) In [63]:

ds ds['number 3'] = ds['number 1'] ds, ds.date.dtype Out[63]: (<xray.Dataset> Dimensions: (date: 10) Coordinates: * date (date) object 10957 10958 10959 10960 10961 10962 10963 10964 ... Data variables: number 1 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 0.6723 ... number 2 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 0.6723 ... number 3 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 0.6723 ..., dtype('O')) In [64]:

ds ds['number 4'] = ds['number 1'][:5] ds, ds.date.dtype Out[64]: (<xray.Dataset> Dimensions: (date: 10) Coordinates: * date (date) int64 10957 10958 10959 10960 10961 10962 10963 10964 ... Data variables: number 1 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 0.6723 ... number 2 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 0.6723 ... number 3 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 0.6723 ... number 4 (date) float64 0.1133 0.5952 0.5467 0.2035 0.2022 nan nan nan ..., dtype('int64')) ```