Question

I got about 2000 sentences about fatalities and i would like to filter them on cause.
To begin i want to start with these:

______ fell (*) ______ 

to the
off the
from the 

where ______ is a group of 1 word, and (*) is to the, off the or from the

I tried

(\w*)fell+\s+to\sthe|off\sthe|from\sthe(\w*)

it returns the “off the” etc but it doesn’t look if the word fell was there.
(the groups probably don’t work neither then.)

So what is wrong, I do use fell+ so fell should be there 1 time right?

Answer 1

I would go with (\\w*)fell\\s[to|off|from\\sthe]\\s*(\\w*)

Here’s a little example:

import java.util.regex.*;
class rtest { 
    static String regex = "(\\w*)fell\\s[to|off|from\\sthe]\\s*(\\w*)";
    static Pattern pattern = Pattern.compile(regex);

    public static void main(String[] args) {
        process("Bob fell off the bike");
        process("Matt fell to the bottom");
        process("I think Terry fell from the beat of a different drum");
    }
    static void process(String text) {
        System.out.println(text);
        String[] tokens = text.split(regex);
        for(String t : tokens) System.out.println(t);
        System.out.println(" ");
    }
}

Results in:

C:\Documents and Settings\glowcoder\My Documents>javac rtest.java

C:\Documents and Settings\glowcoder\My Documents>java rtest
Bob fell off the bike
Bob
 the bike

Matt fell to the bottom
Matt
 the bottom

I think Terry fell from the beat of a different drum
I think Terry
 the beat of a different drum