Median calling ndarray.partition: AttributeError

[Former user]

Calling the median over an ndarray raises an AttributeError:

>>>> import numpy as np
>>>> foo = np.zeros((1,5))
>>>> np.median(foo)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/ester/PhD/pypy2/site-packages/numpy/lib/function_base.py", line 2703, in median
    part = partition(a, (sz - 1) // 2, axis=axis)
  File "/home/ester/PhD/pypy2/site-packages/numpy/core/fromnumeric.py", line 619, in partition
    a.partition(kth, axis=axis, kind=kind, order=order)
AttributeError: 'numpy.ndarray' object has no attribute 'partition'
>>>> np.__version__
'1.8.1'

[Maksym Ganenko]

The same error with numpy.percentile(..)

>>>> import numpy as np
>>>> x = np.array([1,2,3])
>>>> np.percentile(x, 75)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/root/pypy-2.6-linux_x86_64-portable/site-packages/numpy/lib/function_base.py", line 3042, in percentile
    interpolation=interpolation)
  File "/root/pypy-2.6-linux_x86_64-portable/site-packages/numpy/lib/function_base.py", line 2791, in _ureduce
    r = func(a, **kwargs)
  File "/root/pypy-2.6-linux_x86_64-portable/site-packages/numpy/lib/function_base.py", line 3130, in _percentile
    ap.partition(concatenate((indices_below, indices_above)), axis=axis)
AttributeError: 'numpy.ndarray' object has no attribute 'partition'
>>>> np.__version__
'1.9.0'

[Terry Jones]

I've hit the same issue (np.version 1.10.0). I tried using overwrite_input=True just in case it helped, but it doesn't.

It doesn't feel like this should be a hard problem to fix, but I know nothing about numpy/pypy so don't know where to look. Any hints?

[mattip]

I started thinking about this a while ago, and put some analysis here

https://gist.github.com/mattip/ab34268b049b859554ad

Any help would be wonderful

[mattip]

should work now

[mattip]

fixed in af36393