CODEWITHSUNDEEP
javaregex
Zzirconium
Zzirconium

Reputation: 471

How to capture text between uncomplete lines in java

I have got this text (numerical values might change) :

.START_SEQUENCE RANDOM SENTENCE
3.40000
1 2 3 4 some text or not
4 3 8 9
.END_SEQUENCE

I want to get the following text (so basically find everything between .START_SEQUENCE and .END_SEQUENCE, but without neither the end of the START_SEQUENCE line nor the next one)

1 2 3 4 some text or not
4 3 8 9

I have played with Pattern.DOTALL, Pattern.MULTILINE, managed to get rid off things but never ending up on the exact selection I want. I have no clue how to move on.

Here is my last attempt.

final String START_SEQUENCE = "\\.START_SEQUENCE[^\n^\r]*";
final String END_SEQUENCE = "\\.END_SEQUENCE";
Pattern regex = Pattern.compile(START_SEQUENCE+"(.*)"+END_SEQUENCE, Pattern.DOTALL);
Matcher regexMatcher = regex.matcher(emn);
if (regexMatcher.find()) {
    String ResultString = regexMatcher.group(1);
}

Which result is

3.40000
1 2 3 4 some text or not
4 3 8 9

Many thanks in advance !

Upvotes: 1

Views: 113

Answers (3)

aliteralmind
aliteralmind

Reputation: 20163

A non-regex solution: 

import  java.util.ArrayList;
import  java.io.File;
import  java.io.IOException;
import  org.apache.commons.io.FileUtils;
import  org.apache.commons.io.LineIterator;

/**
   <P>{@code java BetweenLineMarkersButSkipFirstXmpl C:\java_code\\xbn\z\xmpl\text\regex\BetweenLineMarkersButSkipFirstXmpl_data.txt}</P>
**/
public class BetweenLineMarkersButSkipFirstXmpl  {
   public static final void main(String[] as_1RqdTxtFilePath)  {
      LineIterator li = null;
      try  {
         li = FileUtils.lineIterator(new File(as_1RqdTxtFilePath[0])); //Throws npx if null
      }  catch(IOException iox)  {
         throw  new RuntimeException("Attempting to open \"" + as_1RqdTxtFilePath[0] + "\"", iox);
      }  catch(RuntimeException rtx)  {
         throw  new RuntimeException("One required parameter: The path to the text file.", rtx);
      }

      String sLS = System.getProperty("line.separator", "\n");

      ArrayList<String> alsItems = new ArrayList<String>();
      boolean bStartMark = false;
      boolean bLine1Skipped = false;
      StringBuilder sdCurrentItem = new StringBuilder();
      while(li.hasNext())  {
         String sLine = li.next().trim();
         if(!bStartMark)  {
            if(sLine.startsWith(".START_SEQUENCE"))  {
               bStartMark = true;
               continue;
            }
            throw  new IllegalStateException("Start mark not found.");
         }  if(!bLine1Skipped)  {
            bLine1Skipped = true;
            continue;
         }  else if(!sLine.equals(".END_SEQUENCE"))  {
            sdCurrentItem.append(sLine).append(sLS);
         }  else  {
            alsItems.add(sdCurrentItem.toString());
            sdCurrentItem.setLength(0);
            bStartMark = false;
            bLine1Skipped = false;
            continue;
         }
      }

      for(String s : alsItems)  {
         System.out.println("----------");
         System.out.print(s);
      }
   }
}

Using this input:

.START_SEQUENCE RANDOM SENTENCE
3.40000
1 2 3 4
4 3 8 9
.END_SEQUENCE
.START_SEQUENCE RANDOM SENTENCE
3.40000
2 3 4 5
3 8 9 10
.END_SEQUENCE

Output:

[C:\java_code\]java BetweenLineMarkersButSkipFirstXmpl C:\java_code\BetweenLineMarkersButSkipFirstXmpl_data.txt
----------
1 2 3 4
4 3 8 9
----------
2 3 4 5
3 8 9 10

Upvotes: 1

nhahtdh
nhahtdh

Reputation: 56809

Use this regex with Pattern.UNIX_LINES flag:

"\\.START_SEQUENCE.*\n.*\n((?:(?!\\.END_SEQUENCE).*\n)*+)\\.END_SEQUENCE"

Explanation

Pattern.UNIX_LINES makes . equivalent to [^\n]. Normally, it is [^\n\r\u0085\u2028\u2029].

Let us break down the regex (to make it easier to read, escape sequences are resolved):

\.START_SEQUENCE.*\n             # Match the .START_SEQUENCE ... line
.*\n                             # Match (and ignore) the next line
((?:(?!\\.END_SEQUENCE).*\n)*+)
\.END_SEQUENCE                   # Match the .END_SEQUENCE line

((?:(?!\\.END_SEQUENCE).*\n)*+) matches the rest of the lines in between and put the result into capturing group 1. Normally, ((?:.*\n)*?) would suffice, but to prevent StackOverflowError on big set of data, I switch to possessive quantifier *+ and a check (?!\\.END_SEQUENCE) is needed so that the repetition can complete without backtracking.

Upvotes: 1

user557597
user557597

Reputation:

Not alot to go on but something like this, and capture group 1 contains data of interest. 

(?-s)\.START_SEQUENCE.*\n.*\n([\S\s]*?)\.END_SEQUENCE

Upvotes: 1

Related Questions