Numpy ndarray object with string and floating point number types in it

Question

I have a file text with data in a following format:

rubbish & 3.97& 3.83& 3.95& 3.83& 3.82

rubbish & 4.92& 4.81& 4.88& 4.81& 4.81

rubbish & 5.90& 5.66& 5.88& 5.66& 5.66

rubbish &--- & 6.05& 6.14& 6.05& 6.05

rubbish & 6.42& 6.26& 6.46& 6.26& 6.26

rubbish &--- & 6.56& 6.63& 6.56& 6.56

And I want to read them into numpy.ndarray object so that numbers are transformed into floating point number object while the --- stay as the string objects. However, the following piece of code creates an expected numpy.array object but everything in there is a string.

import numpy as np

wejscie = open('data.dat', 'r').readlines()

def fun1(x):

  print x
  if x.strip() == '---':

    return str(x)

  else:

    return float(x)

dane = np.array([map(fun1, linijka.split('&')[1:]) for linijka in wejscie])

So is it possible to have numpy.ndarray object containing data of various types?

Try return np.nan instead of return str(x). Or some other float value that will be place holder for the string. You can't mix strings and floats in an array like this. — hpaulj
– hpaulj, Commented Aug 4, 2016 at 16:24

Community · Accepted Answer · 2017-05-23 10:28:43Z

1

The problem isn't with fun1, it's with trying to insert elements of differing types into a numpy array.

Consider the following:

>>> a = numpy.array([1])
>>> numpy.append(a,2)
array([1, 2])
>>> numpy.append(a,'b')
array(['1', 'b'],
  dtype='<U11')

You may find this helpful Store different datatypes in one NumPy array?

edited May 23, 2017 at 10:28

CommunityBot

11 silver badge

answered Aug 4, 2016 at 13:52

dashiell

8124 silver badges11 bronze badges

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

Collectives™ on Stack Overflow

Numpy ndarray object with string and floating point number types in it

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related