Java RegEx - Matching a list of arguments

Question

I am writing a custom parser using regular expressions, but I can't work out how to match functions.

An example of a function in my custom language is:

function int add(int num1, int num2){
  return num1 + num2;
}

My tokenizer uses RegEx to get the next token and remove it from the source code string supplied earlier. This means that when it comes to parsing a function, I can be sure that the code will start with the function statement. I currently have the following expression:

^([\s]*function[\s]+[a-zA-Z][a-zA-Z0-9]*[\s]+[a-zA-Z][a-zA-Z0-9]*[\s]*\(([\s]*[a-zA-Z][a-zA-Z0-9]*[\s]+[a-zA-Z][a-zA-Z0-9]*[\s]*)*\)[\s]*\{.*\}.*)$

It is very long, but it successfully matches these two functions:

function void log(string msg){
  Console.log(msg);
}

and

function int add(int num1 int num2){
  return num1 + num2;
}

I want to be able to split the arguments by a comma.

I could make the comma required after a parameter, but then the last parameter would end with a comma.
I could make the comma optional after a parameter, but then the user would be able to not put a comma in.

I need to be able to only require the comma between parameters, otherwise it will mess up my parser later. How can I edit my expression to look for a comma between arguments?

Thank you very much for your time.

Can you tell me exactly In which function do you want to add comma. — siddhartha jain
– siddhartha jain, Commented Aug 15, 2016 at 0:51
Forget it. You can use a regex to pull individual tokens out of the string, but I wouldn't even try to parse entire constructs using a regex. For one thing, there's no way to use a regex to parse a variable number of parameters and extract all the information. — ajb
– ajb, Commented Aug 15, 2016 at 0:53
I just want to match (int num1, int num2). This is just \(([\s]*[a-zA-Z][a-zA-Z0-9]*[\s]+[a-zA-Z][a-zA-Z0-9]*[\s]*)*\). I need to check that the parameters are split by a comma. — Crazy Redd
– Crazy Redd, Commented Aug 15, 2016 at 1:49

Laurel · Accepted Answer · 2016-08-15 03:54:50Z

1

This regex should work for the (int num1, int num2) part of the string:

(\((?:\s*[^\s,]+\s+[^\s,]+\s*,)*\s*[^\s,]+\s+[^\s,]+\s*\))

It's easier to read when you space it out:

(\(
 (?:      \s*
   [^\s,]+\s+
   [^\s,]+\s*,
 )*       \s*
   [^\s,]+\s+
   [^\s,]+\s*
\))

answered Aug 15, 2016 at 3:54

Laurel

6,23114 gold badges35 silver badges60 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Arthur · Accepted Answer · 2016-08-15 04:41:32Z

0

You could think of it as having 3 different possibilities: zero parameters, one parameter, and more than one parameter. Then just check using the or operator for each different possibility.

One Parameter:

(?:\\w+\\s+\\w+)

More than one parameter:

(?:\\w+\\s+\\w+)(?:\\,\\s+(?:\\w+\\s+\\w+))+

Zero Parameters:

\\s*

When using the or statement (all of the above):

((?:\\w+\\s+\\w+)|(?:\\w+\\s+\\w+)(?:\\,\\s+(?:\\w+\\s+\\w+))+|\\s*)

answered Aug 15, 2016 at 4:41

Arthur

1,2561 gold badge15 silver badges19 bronze badges

Collectives™ on Stack Overflow

Java RegEx - Matching a list of arguments

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related