pattern matching and range checking with regex in java

Question

My application will basically accept user input via the main(String[] args) method. I have been trying to figure a way to do some checking with that input using a patter matcher.

Each valid input consists basically of a single character: (blanks should be ignored)

* 'R' or 'r' - replay
* 'Q' or 'q' - quit
* [1-6] - perform tasks 1 .. 6

If the user input is invalid - which means if we detect anything different than the above, then the application will quit after displaying an error message.

So, all the logic within the main method would be in a single big conditional:

if(Pattern.matches("[RQrq1-6]*", args[0]){

}

I have some difficulty grasping the matches method with regex involved (The above doesn't work) and I would like to avoid having multiple if-else conditions to check each valid input individually.

Can this be achieved with pattern matcher with range checking at the same time?

Thanks for any hints!

Answer 1

I would be tempted not to do this with a regexp, since it'll only match and extract a particular pattern, and not really help with deciding what to do for a particular string (using one regexp).

Instead, I'd check out a command line option library such as Apache Commons CLI. It's much more appropriate for what you require.

Answer 2

I think you need to match each argument passed in. So:

for(int i = 0; i < args.length; i++) {
   if (Pattern.matches("[RQrq1-6]+", args[i]){
     <do stuff>
   }
}

Answer 3

The following class works fine for me (exactly as I'd expect anyway):

public class Test{
  public static void main(String[] args) throws Exception {
    if(java.util.regex.Pattern.matches("[RQrq1-6]*", args[0])){
      System.out.println("OK: "+args[0]);
    } else {
      System.out.println("Not OK: " +args[0]);
    }
  }
}

Example:

burner@firefighter:~$ java Test rr
OK: rr
burner@firefighter:~$ java Test rrRR
OK: rrRR
burner@firefighter:~$ java Test rrRRa
Not OK: rrRRa

Now, you're probably trying to match against all characters on the command line. The command interpreter separates those out into separate strings before launching your application. You could do as Myles suggests, but then you'll process some commands before detecting bad input. You could do this instead:

String myargs = "";
for (String arg : args) {
  myargs += arg;
}
if(java.util.regex.Pattern.matches("[RQrq1-6]*", myargs)){
...

Answer 4

Your regex isn't quite right.

When you write: [RQrq1-6]*

You're saying match zero or more repetitions of any character in the square brackets...

Try this [RQrq1-6] - which means match exactly one.

I wrote a little "spike" program:

import java.util.regex.Pattern;

public class Spike {

public static void main(String[] args) {
 testIt("Q");
 testIt("q");
 testIt("z");
 testIt("Z");
 testIt("R");
 testIt("r");
 testIt("0");
 testIt("1");
 testIt("2");
 testIt("3");
 testIt("4");
 testIt("5");
 testIt("6");
 testIt("7");
 testIt("02");
 testIt("11");
 testIt("17");
 testIt(""); // should fail, used to pass with the question's regex

}

public static void testIt(String sample) {
 System.out.print(sample+" ");
 if (Pattern.matches("[RQrq1-6]*", sample)) {
  System.out.println("match");
 } else {
  System.out.println("fail");
 }
}

}

Output was this:

Q match
q match
z fail
Z fail
R match
r match
0 fail
1 match
2 match
3 match
4 match
5 match
6 match
7 fail
02 fail
11 match
17 fail
 match

Having removed the star in the regex:

Q match
q match
z fail
Z fail
R match
r match
0 fail
1 match
2 match
3 match
4 match
5 match
6 match
7 fail
02 fail
11 fail
17 fail
 fail

How to ignore the blanks

Substitute this version of testit:

public static void testIt(String sample) {
 System.out.print(sample+" ");

 Pattern pattern = Pattern.compile("\\s*([RQrq1-6])\\s*");
 Matcher matcher = pattern.matcher(sample);

 if (matcher.matches()) {
  System.out.println("match '"+matcher.group(1)+"'");
 } else {
  System.out.println("fail");
 }
}

Here are the critical lines of the output:

testIt(" 6");// ignore space tests
testIt(" 7");// ignore space tests
testIt("6 ");// ignore space tests
testIt("7 ");// ignore space tests

    6 match '6'
    7 fail
   6  match '6'
   7  fail

It uses a concept called a "capture group". A capture group is a section of a regex which is assigned a number so that the specific string of symbols that matched can be extracted. We're using a capture group here to pull the symbol out from the optional whitespace.

Answer 5

If I read your question right, you want to call your program like this:

$ myprog 153r34r2q

Then, your program goes through that string of single letter commands and performs them (perhaps to something you pass in in further arguments).

For that, your regex solution is almost OK (use /[RrQq1-6]+/ instead of /[RrQq1-6]*/).

However, I suspect that you actually want to run your program, then accept user commands while the program is running (this assumption is caused by the names of your commands "replay" and "quit", which don't make much sense otherwise). For that, args[] is not suited at all. Instead, you will have to use a main loop which accepts single keystrokes as input. In this case, I wouldn't use a regular expression to test a single character, but use a switch clause for dispatching to the different commands that has a default case to catch invalid input.

Edit: After having read your comment, I propose that you might want to use something along the Command Pattern.