CODEWITHSUNDEEP
pythonnumpy
Bach
Bach

Reputation: 6227

Extra spaces in the representation of numpy arrays of floats

Do you know why the representation of numpy array of integers has no padding:

>>> print array([[1, 2], [3, 4]])
[[1 2]
 [3 4]]

while the numpy array of floats has one extra space before each of the entries?

>>> print array([[1., 2], [3, 4]])
[[ 1.  2.]
 [ 3.  4.]]

(at least in Python 2.7)

I am more interested in the reason/idea behind it, less in the specific implementation details that cause it.

Upvotes: 3

Views: 5011

Answers (3)

Kanav
Kanav

Reputation: 145

Try out below: numpy.set_printoptions(sign=' ')

Upvotes: 2

Revanth
Revanth

Reputation: 11

This will help you:

import numpy as np 
x = np.array(['dev', 'roy', 'kinder'], dtype=np.str)
print("Original Array:")
print(x)
p = np.char.join(" ", x)
print(p)

Upvotes: 0

emesday
emesday

Reputation: 6186

I have looked blame to find out why authors implement as this.

Take a look at the line 584 of https://github.com/numpy/numpy/blame/master/numpy/core/arrayprint.py

self.max_str_len = len(str(int(max_val))) + precision + 2

Because of the magic number 2, the space is added to in front of float values.

Around the line 593, format is determined as

if self.sign:
    format = '%#+'
else:
    format = '%#'
format = format + '%d.%df' % (self.max_str_len, precision)

As your example, max_val is 4.0, and precision is 0. So self.max_str_len will be 3, then format will be %#3.0f

Finally, the value printed is

print '%#3.0f' % 4.
# 4.

Upvotes: 2

Related Questions