CODEWITHSUNDEEP
pythonarrayspython-3.xnumpynumpy-ndarray
Valentin Macé
Valentin Macé

Reputation: 1272

Propagate maximum value in numpy array

I think this is a relatively easy question since I know there is a simple answer in numpy, but I can't find it anymore.

I have a numpy array, say [1, 2, 10, 5, 6, 12, 9, 9, 8]

I want to go through it from beginning to end and to keep only the maximum value seen so far, for example the answer would be [1, 2, 10, 10, 10, 12, 12, 12, 12]

It is very easy to solve with a single loop but some times ago I learned about a numpy function that does the job quickly and that is also compact. I am looking for this method.

Thanks !

Upvotes: 3

Views: 252

Answers (1)

yatu
yatu

Reputation: 88236

You can use numpy.ufunc.accumulate, to accumulate the result of applying a function, maximum in this case, over all elements:

a = np.array([1, 2, 10, 5, 6, 12, 9, 9, 8])

np.maximum.accumulate(a)
# array([ 1,  2, 10, 10, 10, 12, 12, 12, 12], dtype=int32)

Upvotes: 6

Related Questions