How to convert a python list to another data type in numpy?

Question

Write a function that creates a numpy array with the float64 type based on an existing list

import numpy as np

def solution(lst):
    arr = #your code here
    return arr

I try a lot of ways:

import numpy as np

def solution(lst): 
   arr = np.array(lst) 
   arr = lst.astype(np.float64) 
return arr

import numpy as np 
def solution(lst): 
   arr = np.array(lst) 
   arr = np.astype(np.float64) 
return arr

Asked By: badcasemoore

||

Source

Answer 1

A good question will have a minimal test case, and show the errors that you get with various tries.

Define a small list:

In [15]: lst = [1,2,3]

and your two attempts (note the correct indentation for the return):

In [18]: def solution1(lst): 
    ...:    arr = np.array(lst) 
    ...:    arr = lst.astype(np.float64) 
    ...:    return arr
    ...: def solution2(lst): 
    ...:    arr = np.array(lst) 
    ...:    arr = np.astype(np.float64) 
    ...:    return arr
    ...:

First try:

In [19]: solution1(lst)
AttributeError: 'list' object has no attribute 'astype'

astype is a method of a ndarray (as the doc link in my comment shows). list does not have such a method. In python, each class has a defined set of methods. Just because one class has such a method, it does not mean that another has the same.

For the second:

In [21]: solution2(lst)
AttributeError: module 'numpy' has no attribute 'astype'
    
In [22]: np.astype
AttributeError: module 'numpy' has no attribute 'astype'

Again you create an arr, but you try to use astype as np function. You aren’t specifying arr at all in that line.

Make an array from the list:

In [23]: np.array(lst)
Out[23]: array([1, 2, 3])

With this list the dtype is int:

In [24]: np.array(lst).dtype
Out[24]: dtype('int32')

The resulting array has a astype method:

In [25]: np.array(lst).astype(np.float64)
Out[25]: array([1., 2., 3.])

But np.array also accepts a dtype parameter:

In [26]: np.array(lst,dtype=np.float64)
Out[26]: array([1., 2., 3.])

If your list contains floats, you don’t need any further conversion:

In [27]: np.array([1,2,.3])
Out[27]: array([1. , 2. , 0.3])

In [28]: np.array([1,2,.3]).dtype
Out[28]: dtype('float64')

In other words, you should simply have written:

def solution(lst):
    arr = np.array(lst)
    arr = arr.astype(np.float64)   # arr, not lst or np
    return arr

Answered By: hpaulj

How to convert a python list to another data type in numpy?

Question:

Answers: