Parsing Windows tasklist output in Java

Question

I am trying to build an array of processes running on my machine; to do so I have been trying to use the following two commands:

tasklist /fo csv /nh        # For a CSV output
tasklist /nh                # For a non-CSV output

The issue that I am having is that I can not properly parse the output.

First Scenario

I have a line like:

"wininit.exe","584","Services","0","5,248 K"

Which I have attempted to parse using "".split(","), however this fails when it comes to the process memory usage - the comma in the number field willl result in an extra field.

Second Scenario

Without the non-CSV output, I have a line like:

wininit.exe                    584 Services                   0      5,248 K

Which I am attempting to parse using "".split("\\s+") however this one now fails on a process like System Idle Process, or any other process with a space in the executible name.

How can I parse either of these output such that the same split index will always contain the correct data column?

Answer 1

To parse a string, always prefer the most strict formatting. In this case, CSV. In this way, you could process each line with a regular expression containing FIVE groups:

private final Pattern pattern = Pattern
    .compile("\\\"([^\\\"]*)\\\",\\\"([^\\\"]*)\\\",\\\"([^\\\"]*)\\\",\\\"([^\\\"]*)\\\",\\\"([^\\\"]*)\\\"");

private void parseLine(String line) {

    Matcher matcher = pattern.matcher(line);

    if (!matcher.find()) {
        throw new IllegalArgumentException("invalid format");
    }

    String name = matcher.group(1);
    int pid = Integer.parseInt(matcher.group(2));
    String sessionName = matcher.group(3);
    String sessionId = matcher.group(4);
    String memUsage = matcher.group(5);

    System.out.println(name + ":" + pid + ":" + memUsage);
}