python string repsesenting multiple arrays to numpy arrays

Question

I have a string which represents an nD array in the following format:

"[[1, 2, 3, 4, 5], [1, 2, 3, 4, 5], [1, 2, 3, 4, 5], [1, 2, 3, 4, 5]] [[6, 7, 8, 9, 10], [6, 7, 8, 9, 10], [6, 7, 8, 9, 10], [6, 7, 8, 9, 10]] [[11, 12, 13, 14, 15], [11, 12, 13, 14, 15], [11, 12, 13, 14, 15], [11, 12, 13, 14, 15]]"

Now I want to transform this string to a numpy array 4x5x3 (int) (the size could be dynamic based on the number of values within the single and double brackets), something like that:

        /[11, 12, 13, 14, 15]
       / [11, 12, 13, 14, 15]
      /  [11, 12, 13, 14, 15]
     /   [11, 12, 13, 14, 15]    
    /[6, 7, 8, 9, 10]   
   / [6, 7, 8, 9, 10]  
  /  [6, 7, 8, 9, 10] 
 /   [6, 7, 8, 9, 10]                  
[1, 2, 3, 4, 5]   
[1, 2, 3, 4, 5]  
[1, 2, 3, 4, 5] 
[1, 2, 3, 4, 5]

How can this be done efficiently other than looping on every character.

why is the matrix printed in a slanted fashion, also you are missing some commands in your string representation — Devesh Kumar Singh
– Devesh Kumar Singh, Commented Jun 26, 2019 at 11:34
@DeveshKumarSingh I have it slanted in order to show the depth. The double brackets show the size of the 3rd dimension. — ttsesm
– ttsesm, Commented Jun 26, 2019 at 11:37
Why do you have such a string representation in the first place? eval works fine on standard list displays, it does not work with the numpy array prints. — hpaulj
– hpaulj, Commented Jun 26, 2019 at 16:08
well initially my data are in a different form, however I am applying some merging based on this answer stackoverflow.com/a/50169083 which transforms it to the above format. — ttsesm
– ttsesm, Commented Jun 26, 2019 at 17:58

jmd_dk · Accepted Answer · 2019-06-26 11:38:46Z

3

You can simply evaluate the string, transforming it to a tuple of lists of lists. Then convert this to a NumPy array. One trouble is that your string is missing commas after each 2D sub-array (i.e. after ]]). To fix this we just insert these missing commas:

import numpy as np

s = "[[1, 2, 3, 4, 5], [1, 2, 3, 4, 5], [1, 2, 3, 4, 5], [1, 2, 3, 4, 5]] [[6, 7, 8, 9, 10], [6, 7, 8, 9, 10], [6, 7, 8, 9, 10], [6, 7, 8, 9, 10]] [[11, 12, 13, 14, 15], [11, 12, 13, 14, 15], [11, 12, 13, 14, 15], [11, 12, 13, 14, 15]]"
a = np.array(eval(s.replace(']]', ']],')))
print(a)

Note on eval

Using eval is often considered somewhat dangerous, as it can execute arbitrary code. A safer alternative is literal_eval from the ast module:

import ast
a = np.array(ast.literal_eval(s.replace(']]', ']],')))

answered Jun 26, 2019 at 11:38

jmd_dk

13.2k11 gold badges71 silver badges104 bronze badges

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

2 Comments

Paritosh Singh Over a year ago

good approach. though i'd recommend only suggesting the literal_eval version, and talking about eval only in passing if you must, but not as the main answer.

ttsesm Over a year ago

thanks, that works nice. To be honest I was looking on ast.literal_eval but I am still new in python so couldn't figure out how to use it.

Collectives™ on Stack Overflow

python string repsesenting multiple arrays to numpy arrays

1 Answer 1

Note on eval

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Note on eval

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related