How to write bytes to a file in Python 3 without knowing the encoding?

Python Programming

Question or problem about Python programming:

In Python 2.x with ‘file-like’ object:

sys.stdout.write(bytes_)
tempfile.TemporaryFile().write(bytes_)
open('filename', 'wb').write(bytes_)
StringIO().write(bytes_)

How to do the same in Python 3?

How to write equivalent of this Python 2.x code:

def write(file_, bytes_):
    file_.write(bytes_)

Note: sys.stdout is not always semantically a text stream. It might be beneficial to consider it as a stream of bytes sometimes. For example, make encrypted archive of dir/ on remote machine:

tar -c dir/ | gzip | gpg -c | ssh [email protected] 'dd of=dir.tar.gz.gpg'

There is no point to use Unicode in this case.

How to solve the problem:

Solution 1:

It’s a matter of using APIs that operate on bytes, rather than strings.

sys.stdout.buffer.write(bytes_)

As the docs explain, you can also detach the streams, so they’re binary by default.

This accesses the underlying byte buffer.

tempfile.TemporaryFile().write(bytes_)

This is already a byte API.

open('filename', 'wb').write(bytes_)

As you would expect from the ‘b’, this is a byte API.

from io import BytesIO
BytesIO().write(bytes_)

BytesIO is the byte equivalent to StringIO.

EDIT: write will Just Work on any binary file-like object. So the general solution is just to find the right API.

Solution 2:

Specify binary mode, ‘b’, when you open your file:

with open('myfile.txt', 'wb') as w:
    w.write(bytes)

https://docs.python.org/3.3/library/functions.html#open

Hope this helps!