Numpy: Find the maximum length of arrays inside another array

Question

I have a numpy array as the following:

import numpy as np
arr = np.array([np.array([1]),np.array([1,2]),np.array([1,2,3]),np.array([1,3,4,2,4,2])])

I want a nice numpy function, which gives me the maximum length of the arrays inside my arr array. So I need a numpy function, which return for this example 6.

This is possible with iteration, but I am looking for a nicer way, perhaps even without map()

Any function inside tensorflow or keras would also be possible to use.

Your "array" is not an array, it's a Python list, and that particular list has no way of being transformed in any straightforward way into a numpy array (irregular dimension). — Ignacio Vergara Kausel
– Ignacio Vergara Kausel, Commented Sep 3, 2020 at 8:44
max(len(a) for a in arr)? Also, arr can't be a numpy array, only with dtype=object. — Péter Leéh
– Péter Leéh, Commented Sep 3, 2020 at 8:45
Error on my side, I forgot the np.array constructor, I just updated my post — MichaelJanz
– MichaelJanz, Commented Sep 3, 2020 at 8:48
What are the advantages you get by transforming it all to np.arrays? I don't see any. — Péter Leéh
– Péter Leéh, Commented Sep 3, 2020 at 8:49
@MichaelJanz: The answer was wrong. It counted the sum of the elements.Applied to arr = [[9999],[2,2,2,2,2],[1,1]] it would have returned 1. — problemofficer - n.f. Monica
– problemofficer - n.f. Monica, Commented Sep 3, 2020 at 9:12

ansev · Accepted Answer · 2020-09-03 08:50:32Z

3

We could do:

max(map(len, arr))
#6

answered Sep 3, 2020 at 8:50

ansev

31k5 gold badges21 silver badges33 bronze badges

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

3 Comments

Bedir Yilmaz Over a year ago

OP: "This is possible with iteration, but I am looking for a nicer way, perhaps even without map()"

ansev Over a year ago

I think We have to use map or loop here

Bedir Yilmaz Over a year ago

Agreed. But the OP probably wants to know whether an accelerated implementation exists. Possibly an API either with low level language access or with GPU support

Loochie · Accepted Answer · 2022-02-02 12:51:08Z

1

Another simple trick is to use max() function with key argument to return the array with maximum length.

len(max(arr, key = len))

answered Feb 2, 2022 at 12:51

Loochie

2,47215 silver badges20 bronze badges

Comments

MichaelJanz · Accepted Answer · 2020-09-03 09:26:42Z

0

Another way would be to use the keras.preprocessing.sequence.pad_sequences() function.

It pads the sequences to the lenght of the maximum, but in my opinion it creates a high memory usage and depending on the array it might take some time. However, this would be a way without looping:

len(keras.preprocessing.sequence.pad_sequences(arr)[0])

answered Sep 3, 2020 at 9:26

MichaelJanz

1,8252 gold badges10 silver badges24 bronze badges

2 Comments

Ehsan Over a year ago

I expect this to be much slower than simple looping over array.

MichaelJanz Over a year ago

As I said in my answer, yes

Collectives™ on Stack Overflow

Numpy: Find the maximum length of arrays inside another array

3 Answers 3

3 Comments

Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related