How do I convert all of the items in a list to floats? [duplicate]

Question

I have a script which reads a text file, pulls decimal numbers out of it as strings and places them into a list.

So I have this list:

my_list = ['0.49', '0.54', '0.54', '0.55', '0.55', '0.54', '0.55', '0.55', '0.54']

How do I convert each of the values in the list from a string to a float?

I have tried:

for item in my_list:
    float(item)

But this doesn't seem to work for me.

Blaubok

Don't use list as a variable name. – Pentstemon

To elaborate on above comment: using list as a variable name will shadow the built-in list constructor thus not allowing you to use that in the same scope – Heart

Answer 1

[float(i) for i in lst]

to be precise, it creates a new list with float values. Unlike the map approach it will work in py3k.

Libelous

For large array's, I would recommend using numpy: np.array(inp_list, dtype=np.float32). You don't even have to specify if it's a float and just use: np.array(inp_list) – Kenaz

@Thomas Devoogdt : I think you do need to specify the type, otherwise you will get a numpy array of strings – Widget

Please note that you have to assign [float(i) for i in lst] to something. E.g.: new_list = [float(i) for i in lst] – Amaryllidaceous

Why doesn't python have a professional function to perform this operation? – Marchpane

Depending on your security expectations, you might as well just "eval" like eval(my_list.join(", ")) – Ultramontane

Answer 2

map(float, mylist) should do it.

(In Python 3, map ceases to return a list object, so if you want a new list and not just something to iterate over, you either need list(map(float, mylist) - or use SilentGhost's answer which arguably is more pythonic.)

Pentstemon

This is in fact a design sub-optimality in Python 3. – Vaccinia

Answer 3

This would be an other method (without using any loop!):

import numpy as np
list(np.float_(list_name))

Tumbling

Don't need to cast the np.array into the list again if you like to keep it as an np.array =) – Stemma

Answer 4

float(item) do the right thing: it converts its argument to float and and return it, but it doesn't change argument in-place. A simple fix for your code is:

new_list = []
for item in list:
    new_list.append(float(item))

The same code can written shorter using list comprehension: new_list = [float(i) for i in list]

To change list in-place:

for index, item in enumerate(list):
    list[index] = float(item)

BTW, avoid using list for your variables, since it masquerades built-in function with the same name.

Chromonema

Sorry, I did not get what in-place means here. How does it differ from the previous pythonic answer. Doesn't the conversion in the previous answer "[float(i) for i in lst]" retain the original list index – Dashpot

@Dashpot Change original list vs. create a new one. – Chromonema

Answer 5

you can even do this by numpy

import numpy as np
np.array(your_list,dtype=float)

this return np array of your list as float

you also can set 'dtype' as int

Brunhild

But then you are returning a numpy array, not a list. – Kermitkermy

Answer 6

You can use the map() function to convert the list directly to floats:

float_list = map(float, list)

Relativity

Answer 7

You can use numpy to convert a list directly to a floating array or matrix.

    import numpy as np
    list_ex = [1, 0] # This a list
    list_int = np.array(list_ex) # This is a numpy integer array

If you want to convert the integer array to a floating array then add 0. to it

    list_float = np.array(list_ex) + 0. # This is a numpy floating array

Molotov

Answer 8

This is how I would do it.

my_list = ['0.49', '0.54', '0.54', '0.54', '0.54', '0.54', '0.55', '0.54', 
    '0.54', '0.54', '0.55', '0.55', '0.55', '0.54', '0.55', '0.55', '0.54', 
    '0.55', '0.55', '0.54']
print type(my_list[0]) # prints <type 'str'>
my_list = [float(i) for i in my_list]
print type(my_list[0]) # prints <type 'float'>

Tribulation

Answer 9

you can use numpy to avoid looping:

import numpy as np
list(np.array(my_list).astype(float)

Magnolia

Why would you use your approach rather than import numpy as np list(np.float_(list_name)) from the answer above? – Amalee

There appears to be a missing closing parenthesis ). – Spleeny

Answer 10

newlist = list(map(float, mylist))

Measurable

Your answer is correct +1. From python 3.5 you can also do [*map(float, mylist)] to the same effect. Based on https://mcmap.net/q/55080/-getting-a-map-to-return-a-list-in-python-3-x-duplicate – Daleth

Answer 11

import numpy as np
my_list = ['0.49', '0.54', '0.54', '0.54', '0.54', '0.54', '0.55', '0.54', '0.54', '0.54', '0.55', '0.55', '0.55', '0.54', '0.55', '0.55', '0.54', 
'0.55', '0.55', '0.54']
print(type(my_list), type(my_list[0]))   
# <class 'list'> <class 'str'>

which displays the type as a list of strings. You can convert this list to an array of floats simultaneously using numpy:

    my_list = np.array(my_list).astype(np.float)

    print(type(my_list), type(my_list[0]))  
    # <class 'numpy.ndarray'> <class 'numpy.float64'>

Nichy

Answer 12

I had to extract numbers first from a list of float strings:

   df4['sscore'] = df4['simscore'].str.findall('\d+\.\d+')

then each convert to a float:

   ad=[]
   for z in range(len(df4)):
      ad.append([float(i) for i in df4['sscore'][z]])

in the end assign all floats to a dataframe as float64:

   df4['fscore'] = np.array(ad,dtype=float)

Administrator

Answer 13

for i in range(len(list)): list[i]=float(list[i])

Burn

Answer 14

I have solve this problem in my program using:

number_input = float("{:.1f}".format(float(input())))
list.append(number_input)

Ichnite

The OP states that he is reading from a text file. this answer does not apply to what has been asked. – Timmy