Python Sub Process with Tr Command

Question

I am having trouble inserting a tr command into a subprocess.

I have the following:

process = subprocess.Popen('"tr < " + inputFile + " -d '\000' > " + nullFile', shell=True, stdout=subprocess.PIPE)

but keep getting

TypeError: execv() arg 2 must contain only strings

Can anyone see whats going on? It looks like it may be ' and " issues but not sure.

Answer 1

Solved it this way:

command = r"tr -d '\000' < {0:s} > {1:s}".format(inputFile, nullFile)
process = subprocess.Popen(command)
process.wait()

Answer 2

You don't need shell=True , to call tr command:

#!/usr/bin/env python
from subprocess import check_call

with open('input', 'rb', 0) as input_file, \
     open('output', 'wb', 0) as output_file:
    check_call(['tr', '-d', r'\000'], stdin=input_file, stdout=output_file)

Backslash is special inside Python string literals and therefore to pass the backslash, you either need to escape it: '\\\\000' or you should use a raw-string literal: r'\\000' .

You don't need the external process here. You could remove zero bytes from a file using pure Python:

chunk_size = 1 << 15
with open('input', 'rb') as input_file, \
     open('output', 'wb') as output_file:
    while True:
        data = input_file.read(chunk_size)
        if not data: # EOF
            break
        output_file.write(data.replace(b'\0', b''))