Parsing text file into 3 columns

[Vatsa dude]

Hi,
I have a requirement to parse the text file (sample pasted below) and extract the 3 said columns in the file. I am using the Scanner class, but I cannot seem to leverage the "useDelimiter" method to trim the spaces in the text file..

Sample file

64.105.4.90 mail.virtuosoworks.com. ptr
64.105.4.178 mail.imaamd.org. ptr
64.105.4.186 smtp.vernonlaw.com. ptr
64.105.5.25 64-105-5-25.adsl.lbdsl.net. ptr
64.105.5.26 stitch.chipworks.net. ptr
64.105.5.27 studley.chipworks.net. ptr
64.105.5.28 heman.chipworks.net. ptr
64.105.5.29 xena.chipworks.net. ptr
64.105.9.133 MAIL.LINCOLNINDUSTRIAL.COM. ptr
64.105.9.137 DNS1.LINCOLNINDUSTRIAL.COM. ptr

Code below is extracting the 1st column (IP address), but not finding the 2nd and 3rd column. Any help is appreciated. ptr is the 3ed column - its sometimes 1 space after the 2nd column and sometimes multiple spaced after the 2nd column. This file is auto-generated by a mysterious script.

  package com.covad.rdnsmigration;

import java.io.*;
import java.util.Scanner;

public final class ParseTextFile {

	private final File file;

	public static void main(String args[]) throws FileNotFoundException {
		ParseTextFile parser = new ParseTextFile("D:\\depot\\MAIN\\noss\\RDNSmigration\\IpToDNS.txt");
		parser.processLineByLine();
		parser.log("Done");

	}

	public ParseTextFile(String aFile) {
		file = new File(aFile);
	}
	
	public final void processLineByLine() throws FileNotFoundException {
	    Scanner scanner = new Scanner(file);
	    try {
	      //first use a Scanner to get each line
	      while ( scanner.hasNextLine() ){
	        processLine( scanner.nextLine() );
	      }
	    }
	    finally {
	      //ensure the underlying stream is always closed
	      scanner.close();
	    }
	  }

	public void processLine(String aLine){
		    //use a second Scanner to parse the content of each line 
		    Scanner scanner = new Scanner(aLine);
		    scanner.useDelimiter(" ");
		    if ( scanner.hasNext() ){
		      String ipAddress = scanner.next();
		      String rdns = scanner.next();
		      String recordType = scanner.next();
		      log("IP is : " + quote(ipAddress.trim()) + ",  RDNS is : " + quote(rdns.trim()) + ",  and TYPE is : " + quote(recordType.trim()) );
		    }
		    else {
		      log("Empty or invalid line. Unable to process.");
		    }
		    //(no need for finally here, since String is source)
		    scanner.close();
		  }


	private static void log(Object aObject) {
		System.out.println(String.valueOf(aObject));
	}

	private String quote(String aText) {
		String QUOTE = "'";
		return QUOTE + aText + QUOTE;
	}

}
  

[Henry Wong]

//use a second Scanner to parse the content of each line    
Scanner scanner = new Scanner(aLine);   
scanner.useDelimiter(" ");   

The default delimiter is "one or more whitespaces" -- so there is no need to set the delimiter, as it is correct. In fact, the delimiter that you set is exactly one space, and since you mention that there may be more than one space, is actually incorrect.

Henry

[Rob Spoor]

aLine.split("\\s+") is also an option.

[Adeel Ansari]

Rob Prime wrote:aLine.split("\\s+") is also an option.

And this one will be more efficient in terms of performance.

[Rob Spoor]

I wouldn't dare say that without some figures to back it up. Both Scanner and String.split use a java.util.regex.Pattern object. Both these pieces of code create this for each line.
It is likely though that String.split is more efficient since Scanner uses several Pattern objects.

You are right that both are not really efficient. The Pattern can be pulled out of the loop with String.split though:

Pattern pattern = Pattern.compile("\\s+");
for each line
{
    String[] words = pattern.split(line);
}

I've searched the Scanner API but there is no way to reset the Scanner with new input. Therefore, the String.split way will be the more efficient.

[Adeel Ansari]

Rob Prime wrote:I wouldn't dare say that without some figures to back it up...

Actually, I have benchmarked both few months ago, using File IO.. reading ... and then splitting strings based on some token. I tried both Scanner, and String's split. The latter seemed faster.

[Rob Spoor]

Well, then you have some figures to back your claim