Newbie question...

Hi all. I am trying to create a parser that parses something like this:
project ssn, name (relationName);
This is a relational algebra expression.

I have actually made it this far, but I have one slight problem. I want 
the attributes (ssn, name, etc) to become tokens, but without the 
comma... That is, I want the parser to understand that when there is a 
comma, there should be another attribute, but not add the comma to the 
token.

The tokens are defined:
*snip more tokens above*
| <COMMA: ",">
|<ID : ["a"-"z","A"-"Z","_"] ( ["a"-"z","A"-"Z","_","0"-"9"] )*<COMMA> > 
// with comma
|<LASTID: ["a"-"z","A"-"Z","_"] ( ["a"-"z","A"-"Z","_","0"-"9"] )* > 
//without comma(expecting ;)

And the production for the expression:
//----------------------------------
void project() #PROJECT:
{Token t = new Token();}
{// project + attribute name commalist(zero or more(*)) end with 
(relationname) and EOF=;
(t=<PROJECT>) [attributeList()] Relation() {jjtThis.setToken(t.kind, 
t.image);}
}
//-----------------------------------
void attributeList() #ATTRIBUTELIST:
{Token t = new Token();}
{
	(t=<ID>)*{jjtThis.addToken(t.kind, t.image);}
	(t=<LASTID>){jjtThis.addToken(t.kind, t.image);} // optional i.e. can 
project the relation without specific attributes
}
*snipped Relation() as it is irrelevant here*
So...how can I make it so it is parsed the way it is now, but that the 
<COMMA> is not a part of the <ID>s?

Very grateful for advice, have tried alot of different things now.

Regards,
Amund