ufunc.reduceat(a, indices, axis=0, dtype=None, out=None)
Performs a (local) reduce with specified slices over a single axis.
For i in range(len(indices))
, reduceat
computes ufunc.reduce(a[indices[i]:indices[i+1]])
, which becomes the i-th generalized “row” parallel to axis
in the final result (i.e., in a 2-D array, for example, if axis = 0
, it becomes the i-th row, but if axis = 1
, it becomes the i-th column). There are three exceptions to this:
i = len(indices) - 1
(so for the last index), indices[i+1] = a.shape[axis]
.indices[i] >= indices[i + 1]
, the i-th generalized “row” is simply a[indices[i]]
.indices[i] >= len(a)
or indices[i] < 0
, an error is raised.The shape of the output depends on the size of indices
, and may be larger than a
(this happens if len(indices) > a.shape[axis]
).
Parameters: |
a : array_like The array to act on. indices : array_like Paired indices, comma separated (not colon), specifying slices to reduce. axis : int, optional The axis along which to apply the reduceat. dtype : data-type code, optional The type used to represent the intermediate results. Defaults to the data type of the output array if this is provided, or the data type of the input array if no output array is provided. out : ndarray, optional A location into which the result is stored. If not provided a freshly-allocated array is returned. |
---|---|
Returns: |
r : ndarray The reduced values. If |
A descriptive example:
If a
is 1-D, the function ufunc.accumulate(a)
is the same as ufunc.reduceat(a, indices)[::2]
where indices
is range(len(array) - 1)
with a zero placed in every other element: indices = zeros(2 * len(a) - 1)
, indices[1::2] = range(1, len(a))
.
Don’t be fooled by this attribute’s name: reduceat(a)
is not necessarily smaller than a
.
To take the running sum of four successive values:
>>> np.add.reduceat(np.arange(8),[0,4, 1,5, 2,6, 3,7])[::2] array([ 6, 10, 14, 18])
A 2-D example:
>>> x = np.linspace(0, 15, 16).reshape(4,4) >>> x array([[ 0., 1., 2., 3.], [ 4., 5., 6., 7.], [ 8., 9., 10., 11.], [ 12., 13., 14., 15.]])
# reduce such that the result has the following five rows: # [row1 + row2 + row3] # [row4] # [row2] # [row3] # [row1 + row2 + row3 + row4]
>>> np.add.reduceat(x, [0, 3, 1, 2, 0]) array([[ 12., 15., 18., 21.], [ 12., 13., 14., 15.], [ 4., 5., 6., 7.], [ 8., 9., 10., 11.], [ 24., 28., 32., 36.]])
# reduce such that result has the following two columns: # [col1 * col2 * col3, col4]
>>> np.multiply.reduceat(x, [0, 3], 1) array([[ 0., 3.], [ 120., 7.], [ 720., 11.], [ 2184., 15.]])
© 2008–2016 NumPy Developers
Licensed under the NumPy License.
https://docs.scipy.org/doc/numpy-1.10.1/reference/generated/numpy.ufunc.reduceat.html