Equivalent of numpy.add.at for scipy.sparse

Asked May 03 '22 at 15:48

Active May 03 '22 at 16:03

Viewed 47 times

Is there anything like

sparse.add.at(a, indices, b),

or writing a Python for loop is the best we can do when we have duplicates in indices?

edited May 03 '22 at 16:03

hpaulj

asked May 03 '22 at 15:48

Sizhuo Ma

2

Duplicate indices in COO format inputs are summed when converted to CSR. This question needs a [mcve]. – hpaulj May 03 '22 at 16:03
1

More on the sparse handles duplicates, see this recent SO, https://stackoverflow.com/questions/72040697/fastest-add-with-repeated-indices-np-add-at-sparse-csr-matrix – hpaulj May 03 '22 at 16:33

0 Answers0