python: converting an numpy array data type from int64 to int

Question

I am somewhat new to python and I am using python modules in another program (ABAQUS). The question, however, is completely python related.

In the program, I need to create an array of integers. This array will later be used as an input in a function defined in ABAQUS. The problem is to do with the data type of the integers. In the array, the integers have data type 'int64'. However, I am getting the following error when I input the array to the desired function:

"Only INT, FLOAT and DOUBLE supported by the ABAQUS interface (use multiarray with typecode int if standard long is 64 bit)"

I do not need assistance with ABAQUS. If i convert the data type to 'int' in python, that would suffice. I thought that I could simply use the int() function to convert the data type. This did not work. Any suggestions will be highly appreciated. Thank you all.

@Sebastian: That still leaves the data type as int64. It does work for changing the data type from int to float. But not for int64 to int. — Srikanth, Sep 28 '12 at 23:55
That works perfectly well, thank you! I was typing just ".astype(int)". I didn't realize "int" would give the platform dependent data type. Thanks again! — Srikanth, Sep 29 '12 at 00:09
@Srikanth could you provide an answer to your question below and accept it? — bennlich, Feb 11 '15 at 23:45
Possible duplicate of [Converting numpy dtypes to native python types](http://stackoverflow.com/questions/9452775/converting-numpy-dtypes-to-native-python-types) — Hugh Perkins, Aug 31 '16 at 22:45
correct answer, per http://stackoverflow.com/questions/9452775/converting-numpy-dtypes-to-native-python-types is: ``` a.item() ``` — Hugh Perkins, Aug 31 '16 at 22:45

score 34 · Answer 1 · answered Apr 16 '20 at 18:49

34

numpy.ndarray.tolist will do it:

a.tolist()

If your data is a pandas series you can call their tolist wrapper with the same result.

answered Apr 16 '20 at 18:49

bigonazzi

714
7
6

score 25 · Answer 2 · edited Nov 27 '17 at 14:35

25

@J.F. Sebastian's answer:

a.astype(numpy.int32)

edited Nov 27 '17 at 14:35

Community

1
1

answered Feb 13 '15 at 15:05

Srikanth

769
1
8
14

This will work when passing the array to a function. In case, you want to change the type of the array, you need to assign to the array: a = a.astype(...) – Hari Sep 02 '21 at 15:43

score 3 · Answer 3 · answered Mar 23 '18 at 03:05

3

Use the item() method for numpy.int64 object, as Mike T's answer in another similar question explained.

Official documentation is here: https://docs.scipy.org/doc/numpy-dev/reference/generated/numpy.chararray.item.html#numpy.chararray.item

answered Mar 23 '18 at 03:05

Frank Freeman

31
3

score 0 · Answer 4 · edited Jan 19 '22 at 13:33

0

In the end, if you convert an list of int64 values to int with numpy, you will have a numpy value (int64, int32, etc). In case you want a regular int (not numpy int), I found a way which is working. You will convert it to string, and then convert to list! enjoy...

import ast

a = ast.literal_eval(str(a))

edited Jan 19 '22 at 13:33

sarema

695
5
18

answered Jan 19 '22 at 11:33

bardia rafieian

1

score -6 · Answer 5 · edited May 10 '18 at 11:19

-6

If it's a pandas serise, you can first convert it to Dataframe, then use df.to_dict(), then the numpy.int64 will convert to int

In [1]: import pandas as pd

In [2]: import numpy as np

In [3]: df = pd.DataFrame(np.random.randint(5,size=(3,4)),
index=np.arange(3))

In [4]: type(df[0][0])

Out[4]: numpy.int64

In [5]: dict_of_df = df.to_dict()

In [6]: type(dict_of_df[0][0])

Out[6]: int

edited May 10 '18 at 11:19

anothernode

5,100
13
43
62

answered May 10 '18 at 09:30

Charles q

19
5

python: converting an numpy array data type from int64 to int

5 Answers5

Linked