cumprod returns errors

[pwolfram]

The xarray implementation of cumprod returns an assertion error, presumably because of bottleneck, e.g., https://github.com/pydata/xarray/blob/master/xarray/core/ops.py#L333. The error is

└─▪ ./test_cumprod.py 
[ 0.8841785   0.54181236  0.29075258  0.28883015  0.1137352   0.09909713
  0.03570122  0.0304542   0.01578143  0.01496195  0.01442681  0.00980845]
Traceback (most recent call last):
  File "./test_cumprod.py", line 13, in <module>
    foo.cumprod()
  File "/Users/pwolfram/anaconda/lib/python2.7/site-packages/xarray/core/common.py", line 16, in wrapped_func
    skipna=skipna, allow_lazy=True, **kwargs)
  File "/Users/pwolfram/anaconda/lib/python2.7/site-packages/xarray/core/dataarray.py", line 991, in reduce
    var = self.variable.reduce(func, dim, axis, keep_attrs, **kwargs)
  File "/Users/pwolfram/anaconda/lib/python2.7/site-packages/xarray/core/variable.py", line 871, in reduce
    axis=axis, **kwargs)
  File "/Users/pwolfram/anaconda/lib/python2.7/site-packages/xarray/core/ops.py", line 346, in f
    assert using_numpy_nan_func
AssertionError

If bottleneck is uninstalled then a value error is returned:

└─▪ ./test_cumprod.py 
[  2.99508768e-01   2.80142920e-01   1.56389242e-01   1.10791301e-01
   4.58372649e-02   4.10865622e-02   9.91362500e-03   6.76033435e-03
   3.83574249e-03   9.54972340e-04   1.56846616e-04   6.44088547e-05]
Traceback (most recent call last):
  File "./test_cumprod.py", line 13, in <module>
    foo.cumprod()
  File "/Users/pwolfram/anaconda/lib/python2.7/site-packages/xarray/core/common.py", line 16, in wrapped_func
    skipna=skipna, allow_lazy=True, **kwargs)
  File "/Users/pwolfram/anaconda/lib/python2.7/site-packages/xarray/core/dataarray.py", line 991, in reduce
    var = self.variable.reduce(func, dim, axis, keep_attrs, **kwargs)
  File "/Users/pwolfram/anaconda/lib/python2.7/site-packages/xarray/core/variable.py", line 880, in reduce
    return Variable(dims, data, attrs=attrs)
  File "/Users/pwolfram/anaconda/lib/python2.7/site-packages/xarray/core/variable.py", line 213, in __init__
    self._dims = self._parse_dimensions(dims)
  File "/Users/pwolfram/anaconda/lib/python2.7/site-packages/xarray/core/variable.py", line 321, in _parse_dimensions
    % (dims, self.ndim))
ValueError: dimensions () must have the same length as the number of data dimensions, ndim=1

No error occurs if the data array is converted to a numpy array prior to use of cumprod.

This can easily be reproduced by https://gist.github.com/c32f231b773ecc4b0ccf, excerpted below

import numpy as np
import pandas as pd
import xarray as xr                                                                      


data = np.random.rand(4, 3)
locs = ['IA', 'IL', 'IN']
times = pd.date_range('2000-01-01', periods=4)
foo = xr.DataArray(data, coords=[times, locs], dims=['time', 'space'])
print foo.values.cumprod()
foo.cumprod()

[shoyer]

Could you please open a pull request with your branch so we can take a look at what's going on?

The current code will need some adjustments to put cumprod/cumsum in (namely, a variant of reduce that doesn't need to drop dimensions)

[pwolfram]

Thanks @shoyer for the interest. The PR is at #812. I haven't had a ton of time to work on this but it is on my to do list.

[shoyer]

I'm going to close this because it isn't a bug in any merged code. Let's discuss in the PR instead.