Python2.7读取Linux命令dh -h输出遇OSError，求最优读取方法

Hey there! Let's sort out this issue you're hitting when trying to read the output of df -h (I'm guessing you meant df -h instead of dh -h—that's the standard Linux command for checking disk usage, right?) using Python 2.7's subprocess.check_output.

Why the Error Happens #

The root cause here is how check_output handles command arguments in Python 2.7 by default. If you pass a single string like "df -h" without setting shell=True, Python tries to look for an executable file literally named df -h (with the space included) instead of running the df command with the -h flag. That's why you get the "No such file or directory" error.

The Best Ways to Read Linux Command Output #

Let's go through the optimal approaches, ordered by best practice:

1. Split the Command Into a List (Recommended, Most Secure)

This is the safest and most efficient method because it avoids spawning a shell, eliminating the risk of shell injection attacks (critical if your command includes user-provided input). Pass the command and its arguments as separate elements in a list:

import subprocess

try:
    # Split command and args into a list
    raw_output = subprocess.check_output(['df', '-h'])
    # Convert bytes to string (Python 2.7 returns bytes; strip removes extra newlines)
    output = raw_output.strip()
    print(output)
except subprocess.CalledProcessError as e:
    # Handle cases where the command exits with a non-zero code
    print(f"Command failed with exit code {e.returncode}: {e.output}")
except OSError as e:
    # Handle cases where the command itself isn't found
    print(f"Could not execute command: {e}")

2. Use shell=True (Only When Necessary)

If your command relies on shell features like pipes (|), wildcards (*), or environment variable expansion, you can use shell=True. But be cautious—never use this with untrusted input, as it can lead to security vulnerabilities:

import subprocess

try:
    # Use shell=True to run the command through the system shell
    raw_output = subprocess.check_output('df -h', shell=True)
    output = raw_output.strip()
    print(output)
except subprocess.CalledProcessError as e:
    print(f"Command failed: {e}")
except OSError as e:
    print(f"Execution error: {e}")

Bonus: Handling Encoding (If Needed) #

In Python 2.7, check_output returns a byte string. If you need to work with Unicode, decode it using the appropriate encoding for your system (usually utf-8):

output_str = raw_output.decode('utf-8').strip()

Why the List Method Is Better #

• Security: No shell means no risk of malicious input being executed as part of the command.
• Efficiency: Avoids the overhead of spawning a separate shell process.
• Clarity: Explicitly separates the command from its arguments, making your code easier to read and debug.

内容的提问来源于stack exchange，提问作者preethy tulpi