MATLAB arrayfun alternative in Python

Question

In MATLAB, I execute something like this:

z = diag(c)*D*x;
idxN = z<0; idxP = z>=0; % logical indexing
y1 = sum(-z(idxN))+sum(arrayfun(@(x) log(exp(x)+1),z(idxN)));

I'm trying to formulate a Python one:

z = np.diagflat(c).dot(D).dot(x)
idxN = z<0, idxP = z>=0
y1 = np.sum(-z[idxN])+np.sum(math.log(np.exp(x)+1),z[idxN]);

The above snippet doesn't work as I'm guessing I've to vectorize somewhere?

Thanks for your inputs.

arrayfun isn't vectorization, it's shorthand for a loop. Just use a loop, or rewrite your code to use vectorization. — Wolfie
– Wolfie, Commented Jun 20, 2018 at 8:09

Sardar Usama · Accepted Answer · 2018-06-20 15:46:11Z

1

arrayfun is a one-line wrapper for a loop in MATLAB. In Python, you can do it like this:

y1 = np.sum(-z[idxN]) + np.sum([math.log(np.exp(x)+1) for x in z[idxN]])

answered Jun 20, 2018 at 15:46

Sardar Usama

19.7k9 gold badges39 silver badges60 bronze badges

Sign up to request clarification or add additional context in comments.

1 Answer 1