Why does numpy convert int to string and not string to int in a mixed array?

Question

I have this code:

import numpy as np

def printMe(myData):
    print(myData)
    print(type(myData))
    print("length = " + str(len(myData)))
    for x in myData:
        print(x)
        print(type(x))

if __name__ == '__main__':
    myOtherData = np.array([123,"456"])
    printMe(myOtherData)

The output is

['123' '456']
<class 'numpy.ndarray'>
length = 2
123
<class 'numpy.str_'>
456
<class 'numpy.str_'>

Why does 123 get converted to a string rather than converting the '456' to an integer?

The authors of numpy designed it that way. Probably because int to str conversion can't fail unlike the other way. — Michael Butscher
– Michael Butscher, Commented Jul 17, 2021 at 13:13

Mutaz-MSFT · Accepted Answer · 2021-07-17 13:13:09Z

1

If dtype is not specified in the call np.array then the type will be determined as the minimum type required to hold the objects in the sequence. In your example that will be a string.

Otherwise if you use np.array([123,'456'], dtype=int), then it will try and force int type for all sequence members, this however will fail with value error if one of the members is not suitable to be converted to the desired dtype.

answered Jul 17, 2021 at 13:13

Mutaz-MSFT

8066 silver badges20 bronze badges

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

2 Comments

nicomp Over a year ago

Makes sense. It's different in np.genfromtxt() , where the non-numerics are converted to nan and the numerics are preserved.

hpaulj Over a year ago

The default dtype for genfromtxt is float. loadtxt raises an error if it can't make the conversion. Both start with string data (from the file) and choose a general numeric dtype.

Collectives™ on Stack Overflow

Why does numpy convert int to string and not string to int in a mixed array?

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related