CODEWITHSUNDEEP
pythonfilenumpyio
ramailo sathi
ramailo sathi

Reputation: 355

numpy array input from file

I am reading lines from file in a list:

import numpy as np

lines = tuple(open('values.txt','r'))
x = np.array([line for line in lines])

values.txt looks like:

[1,0,1,0],
[1,0,0,0]

It throws an error:

valueError: invalid literal for float()

However, if I just assign the list to x, it works just fine.

How to take input from file in a numpy array?

Upvotes: 1

Views: 1442

Answers (3)

Mike Müller
Mike Müller

Reputation: 85462

You can read the whole file at once and just add the outer [] before applying literal_eval() to all lines:

from ast import literal_eval

with open('values.txt') as fobj:
    x = np.array(literal_eval('[{}]'.format(fobj.read())))

Upvotes: 0

jonatan
jonatan

Reputation: 135

import numpy as np
import ast

lines = open('values.txt','r')
x = np.array([ast.literal_eval(line.strip(',\n')) for line in lines])

Upvotes: 1

Alex Alifimoff
Alex Alifimoff

Reputation: 1849

lines = open('values.txt', 'r')
x = np.array( [  map(float, (l[l.find("[")+1 : l.find("]")].split(",")))
     for l in lines ] )
print x

A brief explanation:

This takes each line in your file, finds the brackets on each side, and takes the string within the brackets. We then split that string into an array using commas as the delimiter. Then, we have an array of strings so we map the float function onto each element, turning it into a floating point number. Then we use the standard list comprehension to do this to every line.

Upvotes: 1

Related Questions