Converting a numpy.ndarray to string

Use ndarray.tostring -

my_string_numpy_array.tostring()

Sample output -

In [176]: my_string_numpy_array.tostring()Out[176]: 'My name is Aman Raparia'

The right answer for numpy is Divakar's ndarray.tostring.

An alternative is to use chr on each array element and join together (for a non numpy array for example):

>>> ''.join([chr(e) for e in my_string_numpy_array])'My name is Aman Raparia'

Now some years later, np.fromstring() is deprecated:

my_string_numpy_array = np.fromstring(my_string, dtype=np.uint8):1: DeprecationWarning: The binarymode of fromstring is deprecated, as it behaves surprisingly onunicode inputs. Use frombuffer instead

Use np.frombuffer() instead. And switch back to the input with np.array2string().

import numpy as npmy_string = b"My name is Aman Raparia"my_string_numpy_array = np.frombuffer(my_string, dtype=np.uint8)np.array2string(my_string_numpy_array, formatter={'int':lambda x: chr(x).encode()}, separator='').strip('[]').encode()

Out: b'My name is Aman Raparia'

You can drop the .encode() to get string directly. It is just added to get to the original input again, and np.frombuffer() requires a byte formatted buffer.

If you have more than one item as output and then need to remove the []-brackets, see np.array2string not removing brackets around an array.