[mohit]

Hi all,

I want to do some reverse parsing on string using regular expression.
see e.g.
string is : "cancer of lung as cancer example to get cancer."

i want to write a regular expression which can return me
"cancer" only in case "cancer" following "of lung".

can We parse it using single regular expression?

Regards,
Mohit

[mohit]

I got it.

i can use this

[a-zA-Z]+(?=( )of( )lung)

but take the below example

"lung cancer of lung is called as something of lung "

now my requirement is i want "cancer lung" and "something lung" through one regular expression.



any help.

[UnrealEd]

you mean like this:

Code:

([a-zA-Z]+\s*(?=lung))

[mohit]

Thanks for your quick reply.
but
it is returning me only "of" instead of either side text.
i mean i need "cancer lung" from "cancer of lung".

any guess.
Thanks

[UnrealEd]

ow, sorry, I thought the string was "cancer lung". Well then!

This should do:

Code:

(([a-zA-Z]+)(?:\sof)?\s(?=lung))

this will result in ('cancer of lung', 'cancer') or ('something lung', 'something')

[mohit]

Sorry to say but i want "cancer lung" from "cancer of lung".

your expression is returning me cancer of or something of

[UnrealEd]

Did you take a look at the Capturing Groups Tutorial? The second matched group will contain "cancer" or "something". If all strings you want to find end with "lung", you can simply select the second matched group, and add "lung" to it, like so:

Pattern p = Pattern.compile("(([a-zA-Z]+)(?:\sof)?\s(?=lung))");
Matcher m = Pattern.matcher (
"lung cancer of lung is called as something of lung ");
System.out.println (m.group(1) + " lung"); // it might be 2, depending on how Java numbers the groups 
 

[mohit]

Thanks,

but i have done it in something different way.
i used this

([a-zA-Z]+/[A-Z]+) (of/[A-Z]+) ([a-zA-Z]+/[A-Z]+)

then use replace it with $1 $3

Code:

String regex ="([a-zA-Z]+/[A-Z]+) (of/[A-Z]+) ([a-zA-Z]+/[A-Z]+)";
Pattern pattern=Pattern.compile(regex);
Matcher matcher= pattern.matcher(string);
while(matcher.find()){
		String str="";
                str= string.substring(matcher.start(),matcher.end());
                 str = pattern.matcher(str).replaceAll("$1 $3");
}

it worked for me.

thanks for contribution

[UnrealEd]

that will match any 3 word sentence, not just "cancer of lung"

Anyway, I'm glad you got it solved

[hem1234]

Originally Posted by mohit

Thanks,

but i have done it in something different way.
i used this

([a-zA-Z]+/[A-Z]+) (of/[A-Z]+) ([a-zA-Z]+/[A-Z]+)

then use replace it with $1 $3

Code:

String regex ="([a-zA-Z]+/[A-Z]+) (of/[A-Z]+) ([a-zA-Z]+/[A-Z]+)";
Pattern pattern=Pattern.compile(regex);
Matcher matcher= pattern.matcher(string);
while(matcher.find()){
		String str="";
                str= string.substring(matcher.start(),matcher.end());
                 str = pattern.matcher(str).replaceAll("$1 $3");
}

it worked for me.

thanks for contribution

Thanks for the code