Trying to convert int arrays to string arrays in numpy
In [66]: a=array([0,33,4444522])
In [67]: a.astype(str)
Out[67]:
array(['0', '3', '4'], dtype='|S1')Not what I intended
In [68]: a.astype('S10')
Out[68]:
array(['0', '33', '4444522'], dtype='|S10')This works but I had to know 10 was big enough to hold my longest string. Is there a way of doing this easily without knowing ahead of time what size string you need? It seems a little dangerous that it just quietly truncates your string without throwing an error.
15 Answers
Again, this can be solved in pure Python:
>>> map(str, [0,33,4444522])
['0', '33', '4444522']Or if you need to convert back and forth:
>>> a = np.array([0,33,4444522])
>>> np.array(map(str, a))
array(['0', '33', '4444522'], dtype='|S7')You can stay in numpy, doing
np.char.mod('%d', a)This is twice faster than map or list comprehensions for 10 elements, four times faster for 100. This and other string operations are documented here.
Use arr.astype(str), as int to str conversion is now supported by numpy with the desired outcome:
import numpy as np
a = np.array([0,33,4444522])
res = a.astype(str)
print(res)
array(['0', '33', '4444522'], dtype='<U11')You can find the smallest sufficient width like so:
In [3]: max(len(str(x)) for x in [0,33,4444522])
Out[3]: 7Alternatively, just construct the ndarray from a list of strings:
In [7]: np.array([str(x) for x in [0,33,4444522]])
Out[7]:
array(['0', '33', '4444522'], dtype='|S7')or, using map():
In [8]: np.array(map(str, [0,33,4444522]))
Out[8]:
array(['0', '33', '4444522'], dtype='|S7')np.apply_along_axis(lambda y: [str(i) for i in y], 0, x)
Example
>>> import numpy as np
>>> x = np.array([-1]*10+[0]*10+[1]*10)
array([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1])
>>> np.apply_along_axis(lambda y: [str(i) for i in y], 0, x).tolist()
['-1', '-1', '-1', '-1', '-1', '-1', '-1', '-1', '-1', '-1', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '1', '1', '1', '1', '1', '1', '1', '1', '1', '1']
