how to read subprocess stdout using a different separator from line break

Question

I think I can use from J.F. Sebastian's answer in another thread to read from a subprocess line by line, using a line buffer

p = Popen(["command", "args"], stdout=PIPE, bufsize=1)
with p.stdout:
    for line in iter(p.stdout.readline, b''):
        print(line, end='\n')

The problem I'm trying to solve is my output doesn't have line breaks, but uses special characters (such as "#") to indicate termination, which I want to treat as line breaks when printing.

You'll have to buffer the output yourself. Might be helpful to post some sample output from your process, and how you'd like to actually display it. — Wayne Werner, Apr 18 '16 at 23:18
1- As the answer that you've linked indicates, you could use `for line in p.stdout` on Python 3 (as hinted by the `print()` function in your code) instead of `iter(..)` 2- To match the stream against a regex, [`python-streamexpect` package](https://github.com/digidotcom/python-streamexpect/issues/3) provides an interface similar to `pexpect`. Related: [Reading a file with a specified delimiter for newline](http://stackoverflow.com/q/16260061/4279) (`p.stdout` is file-like) — jfs, Apr 19 '16 at 15:02

score 3 · Accepted Answer · answered Apr 19 '16 at 00:02

You could read a char at a time checking for delimiters:

p = Popen(["command", "args"], stdout=PIPE, universal_newlines=True)
tmp = ""
delims = {"#", "!"}
for ch in iter(lambda: p.stdout.read(1), ""):
    if ch in delims:
        print(tmp)
        tmp = ""
    else:
        tmp += ch

Not the prettiest solution but if you have multiple delimiters I don't see many other ways to do it.

how to read subprocess stdout using a different separator from line break

1 Answers1