Use Regex to Group on Sub Strings

[Steve Dyke]

This is an example of one of my strings:

SEAT AY-AAR CTR IAT IVS TPL

This is what I would like my result to be:

CTR IAT IVS TPL

My code:

public final static String QMS_ALLOWED_CHARS_CSS_DESC = "[^(L/H|R/H|IAT|CTR|IVS|TPL)\040]";

myString.trim().replaceAll(GenericPrams.QMS_ALLOWED_CHARS_CSS_DESC, "");

Results I am getting:
SAT AAAR CTR IAT IVS TPL

[Carey Brown]

What's enclosed in brackets  [....] are a list of characters, or [^....] a list of NOT characters. You seem to be mixing strings (r/h|l/h|...) inside a list of characters.

Without knowing more about your input patterns it would be hard to recommend a regex. For instance, does input always start with "SEAT " ?

[Carey Brown]

Somebody may have a trickier way to do this with one line but this is what I came up with that may suit your needs.

import java.util.regex.Matcher;
import java.util.regex.Pattern;

public class RegexTickets
{
	// each 3 letter pattern must be preceded by a space
	private static final Pattern PATTERN =
			Pattern.compile( "( (L/H|R/H|IAT|CTR|IVS|TPL))" );
	
	public static void main( String[] args )
	{
		String			ticket = "SEAT AY-AAR CTR IAT IVS TPL";
		System.out.println( applyPattern( ticket ) );
	}
	
	private static String applyPattern( String s )
	{
		Matcher			m = PATTERN.matcher( s );
		StringBuilder	sb = new StringBuilder();
		while( m.find() )
			sb.append( m.group(1) );
		sb.delete( 0, 1 ); // remove leading space
		return sb.toString();
	}
}

[Ron McLeod]

There is probably a pure REGEX way to solve the problem, but this is how I would tackle it:

import java.util.stream.Collectors;
import java.util.stream.Stream;

public class JR33246 {
	private static final String SPACE = " ";

	public static void main(String[] args) {
		String codes = "SEAT AY-AAR CTR IAT WXYZ IVS TPL";
		String regex = "(L/H|R/H|IAT|CTR|IVS|TPL)";

		String cleaned = Stream.of(codes.split(SPACE))
				.filter(w -> w.matches(regex))
				.collect(Collectors.joining(SPACE));

		System.out.println(cleaned);
	}
}

CTR IAT IVS TPL
Edited to fix code typo

[salvin francis]

I think you meant :

private static final String SPACE = " ";

you missed the space in SPACE.

[Ron McLeod]

salvin francis wrote:you missed the space in SPACE.

Thanks for spotting that - fixing it now  

[salvin francis]

Unless I have completely misunderstood OP's question, here's a simple regex I wrote:

public class Main {
    public static void main(String ...salvin) {
        String codes = "SEAT AY-AAR CTR IAT WXYZ IVS TPL";
        String regex = "SEAT\\s*|AY-AAR\\s*|WXYZ\\s*";
        System.out.println(codes.replaceAll(regex, ""));
    }
}

[Carey Brown]

I think that the OP knows the set of patterns he wants to accept and wants to reject everything else. Therefore we don't know if "WXYZ" will appear, or "ABCD", but if it's not one of the accepted patterns then don't include it. I can't think of a way to do this with replaceAll().

[Steve Dyke]

Carey Brown wrote:I think that the OP knows the set of patterns he wants to accept and wants to reject everything else. Therefore we don't know if "WXYZ" will appear, or "ABCD", but if it's not one of the accepted patterns then don't include it. I can't think of a way to do this with replaceAll().

Yes, you have correctly interrupted my challenge. I will try your previous solution. Thanks.

[Steve Dyke]

Ron McLeod wrote:There is probably a pure REGEX way to solve the problem, but this is how I would tackle it:

import java.util.stream.Collectors;
import java.util.stream.Stream;

public class JR33246 {
	private static final String SPACE = " ";

	public static void main(String[] args) {
		String codes = "SEAT AY-AAR CTR IAT WXYZ IVS TPL";
		String regex = "(L/H|R/H|IAT|CTR|IVS|TPL)";

		String cleaned = Stream.of(codes.split(SPACE))
				.filter(w -> w.matches(regex))
				.collect(Collectors.joining(SPACE));

		System.out.println(cleaned);
	}
}

CTR IAT IVS TPL
Edited to fix code typo

This would solve my challenge but we are not using Java 8 yet. Thanks.

[Carey Brown]

I like Ron's example because it's clean and robust (but is Java-8).
My example suffers from matching things that are more than 3 characters, e.g. "SEAT AY-AAR CTR TPLQ XYZ". It will see "TPL" as a match even though it is "TPLQ", which is not a match.
A hybrid solution would work best, I think.

	private static String applyPattern( String s )
	{
		String regex = "(L/H|R/H|IAT|CTR|IVS|TPL)";

		List<String>	list = new ArrayList<>();
		String[]		words = s.split( " +" );
		for( String word : words )
			if( word.matches( regex )  )
				list.add( word );
		return String.join( " ", list );
	}