put(x, indices, y, axis=-1, ufunc=<ufunc 'add'>)¶
A kind of inverse mapping of np.take
In a simple, the operation can be thought as:
x[indices] += y
with the exception that all entries of y are used instead of just the first occurence corresponding to a particular element. That is, the results are accumulated, and the accumulation function can be changed by providing ufunc. For instance, np.multiply corresponds to:
x[indices] *= y
Whereas np.take picks indices along an axis and returns the resulting array, put similarly picks indices along an axis but accumulates the given values to those entries.