gyshi commented on a change in pull request #16257: [Numpy] add numpy op
bitwise_xor, hsplit, moveaxis, rot90
URL: https://github.com/apache/incubator-mxnet/pull/16257#discussion_r329308486
##########
File path: python/mxnet/ndarray/numpy/_op.py
##########
@@ -2012,6 +2012,121 @@ def split(ary, indices_or_sections, axis=0):
# pylint: enable=redefined-outer-name
+# pylint: disable=redefined-outer-name
+@set_module('mxnet.ndarray.numpy')
+def hsplit(ary, indices_or_sections):
+ """Split an array into multiple sub-arrays horizontally (column-wise).
+
+ This is equivalent to ``split`` with ``axis=0`` if ``ary`` has one
+ dimension, and otherwise that with ``axis=1``.
+
+ Parameters
+ ----------
+ ary : ndarray
+ Array to be divided into sub-arrays.
+ indices_or_sections : int, list of ints or tuple of ints.
+ If `indices_or_sections` is an integer, N, the array will be divided
+ into N equal arrays along `axis`. If such a split is not possible,
+ an error is raised.
+
+ If `indices_or_sections` is a list of sorted integers, the entries
+ indicate where along `axis` the array is split.
+
+ If an index exceeds the dimension of the array along `axis`,
+ it will raises errors. so index must less than or euqal to
+ the dimension of the array along axis.
+
+ Returns
+ -------
+ sub-arrays : list of ndarrays
+ A list of sub-arrays.
+
+ Notes
+ ------
+ - If `indices_or_sections` is given as an integer, but a split
+ does not result in equal division.It will raises ValueErrors.
+
+ - If indices_or_sections is an integer, and the number is 1, it will
+ raises an error. Because single output from split is not supported yet...
+
+ See Also
+ --------
+ split : Split an array into multiple sub-arrays of equal size.
+
+ Examples
+ --------
+ >>> x = np.arange(16.0).reshape(4, 4)
+ >>> x
+ array([[ 0., 1., 2., 3.],
+ [ 4., 5., 6., 7.],
+ [ 8., 9., 10., 11.],
+ [12., 13., 14., 15.]])
+ >>> np.hsplit(x, 2)
+ [array([[ 0., 1.],
+ [ 4., 5.],
+ [ 8., 9.],
+ [12., 13.]]),
+ array([[ 2., 3.],
+ [ 6., 7.],
+ [10., 11.],
+ [14., 15.]])]
+ >>> np.hsplit(x, [3, 6])
+ [array([[ 0., 1., 2.],
+ [ 4., 5., 6.],
+ [ 8., 9., 10.],
+ [12., 13., 14.]]),
+ array([[ 3.],
+ [ 7.],
+ [11.],
+ [15.]]),
+ array([], shape=(4, 0), dtype=float32)]
+
+ With a higher dimensional array the split is still along the second axis.
+
+ >>> x = np.arange(8.0).reshape(2, 2, 2)
+ >>> x
+ array([[[ 0., 1.],
+ [ 2., 3.]],
+ [[ 4., 5.],
+ [ 6., 7.]]])
+ >>> np.hsplit(x, 2)
+ [array([[[ 0., 1.]],
+ [[ 4., 5.]]]),
+ array([[[ 2., 3.]],
+ [[ 6., 7.]]])]
+
+ If ``ary`` has one dimension, 'axis' = 0.
+ >>> x = np.arange(4)
+ array([0., 1., 2., 3.])
+ >>> np.hsplit(x, 2)
+ [array([0., 1.]), array([2., 3.])]
+
+ If you want to produce an empty sub-array, you can see an example.
+ >>> np.hsplit(x, [2, 2])
+ [array([0., 1.]), array([], dtype=float32), array([2., 3.])]
+ """
+ indices = []
+ axis = 1
+ if (len(ary.shape) == 1):
+ axis = 0
+ axis_size = ary.shape[axis]
+ if isinstance(indices_or_sections, int):
+ sections = indices_or_sections
+ if axis_size % sections:
+ raise ValueError('array hsplit does not result in an equal
division')
+ section_size = int(axis_size / sections)
+ indices = [i * section_size for i in range(sections)]
+ elif isinstance(indices_or_sections, (list, set, tuple)):
+ indices = [0] + list(indices_or_sections)
+ else:
+ raise ValueError('indices_or_sections must either int or tuple of
ints')
+ ret = _npi.hsplit(ary, indices, axis, False)
+ if not isinstance(ret, list):
+ raise NotImplementedError('single output from hsplit is not supported
yet...')
+ return ret
Review comment:
I know what you mean, but in numpy, when your input data is 1 dim, the axis
will be 0, if ndim > 1, axis will be 1.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
