In Python how to find max length of inner list in a 2D list?

Question

I often create a list of lists with varying length of inner list, for e.g. to represent a bunch of sentences of varying length

[['Hello', 'world'], ['how','are','you'], ['have','a','good','day']]

I need these to be converted to numpy matrix. To make all inner lists of same length I pad a dummy element at the end and make all inner lists equal the maximum length.

Is there any compact way of finding the max length of inner list?

I usually do it by writing a for loop and keeping track of the max length but I do it so often that I feel the need for a better way.

How is this `numpy`? the list contains strings, not interger — U13-Forward, Jul 21 '18 at 07:16
Pad without explicitly calculating longest: `np.array(list(itertools.zip_longest(*alist, fillvalue='dummy'))).T` — hpaulj, Jul 21 '18 at 07:27
@U9-Forward sorry I forgot to mention I use word embedding like GloVe to replace each word with a vector, that's where I use numpy. — Rakesh K, Jul 21 '18 at 12:00

Netwave · Accepted Answer · 2018-07-21T07:10:07.437

15

Using max function:

max(your_list, key=len)

You will get the longest list, if you need the actual length just use len again:

len(max(your_list, key=len))

Here you have the live example

edited Jul 21 '18 at 07:10

answered Jul 21 '18 at 07:09

Netwave

40,134
6
50
93

score 4 · Answer 2 · answered Jul 21 '18 at 07:11

4

Using map and max, you find the max length of the sub lists easily

>>> max(map(len, lst))
4

answered Jul 21 '18 at 07:11

Sunitha

11,777
2
20
23

score 2 · Answer 3 · answered Jul 21 '18 at 07:07

2

you can do it like the following:

max([len(item) for item in A])

answered Jul 21 '18 at 07:07

OmG

18,337
10
57
90

better as a generator than using an intermediate list – Netwave Jul 21 '18 at 07:07

score 2 · Answer 4 · answered Jul 21 '18 at 07:14

2

And order it:

>>> A=[['Hello', 'world'], ['how','are','you'], ['have','a','good','day']]
>>> sorted(A,key=len)
[['Hello', 'world'], ['how', 'are', 'you'], ['have', 'a', 'good', 'day']]
>>>

answered Jul 21 '18 at 07:14

U13-Forward

69,221
14
89
114

Guillaume Jacquenot · Answer 5 · 2018-07-21T07:21:35.357

-1

To convert your list of lists in a numpy array, you can try the following snippet. The fourth line pads your data and the fifth line creates the numpy array.

import numpy as np
data=[['Hello', 'world'], ['how','are','you'], ['have','a','good','day']]
max_length = max(map(len, data))
[e1.extend(['']*(max_length - len(e1))) for e1 in data]
np_data=np.array(data)

edited Jul 21 '18 at 07:21

answered Jul 21 '18 at 07:16

Guillaume Jacquenot

11,217
6
43
49

3

don't use list-comprehensions for side-effects. – juanpa.arrivillaga Jul 21 '18 at 07:19

In Python how to find max length of inner list in a 2D list?

5 Answers5