Question

I have a small interrogation concerning my grammar.
I want to parse strings, like the following :

 "(ICOM LIKE '%bridge%' or ICOM LIKE '%Munich%')"

I ended up with the following grammar (a bit more complex than needed I know) :

// Aiming at parsing a complete BQS formed Query

grammar Logic;

options {
    output=AST;
}

tokens {
  NOT_LIKE;
}

/*------------------------------------------------------------------
 * PARSER RULES
 *------------------------------------------------------------------*/
 // precedence order is (low to high): or, and, not, [comp_op, geo_op, rel_geo_op, like, not like, exists], ()
 parse  
    : expression EOF -> expression
    ; // ommit the EOF token

 expression
    : query
    ;       

 query  
    : term (OR^ term)*    // make `or` the root
    ;

 term   
    : factor (AND^ factor)*
    ;

 factor
  :  (notexp -> notexp) ( NOT LIKE e=notexp  -> ^(NOT_LIKE $factor $e))?
  ;

 notexp
  :  NOT^ like
  |  like
  ;

 like // this one has to be completed (a lot)
    : atom (LIKE^ atom)*
    ;


 atom   
    : ID 
    | | '(' expression ')' -> expression
    ;

/*------------------------------------------------------------------
 * LEXER RULES
 *------------------------------------------------------------------*/
// GENERAL OPERATORS: 
//NOTLIKE   :   'notlike' | 'NOTLIKE'; // whitespaces have been removed
LIKE    :   'like' | 'LIKE';

OR          :   'or' | 'OR';
AND         :   'and' | 'AND';
NOT         :   'not' | 'NOT';

//ELEMENTS 
CONSTANT_EXPRESSION : DATE | NUMBER | QUOTED_STRING;    
ID          :   (CHARACTER|DIGIT)+; 

WHITESPACE : ( '\t' | ' ' | '\r' | '\n'| '\u000C' )+    { $channel = HIDDEN; } ;

fragment DATE       :   '\'' YEAR '/' MONTH '/' DAY (' ' HOUR ':' MINUTE ':' SECOND)? '\'';

fragment QUOTED_STRING :    '\'' (CHARACTER)+ '\'' ; 

//UNITS
fragment CHARACTER :    ('a'..'z' | 'A'..'Z'|'.'|'\''|'%'); // FIXME: Careful, should be all ASCII
fragment DIGIT  :   '0'..'9' ;
fragment DIGIT_SEQ  :(DIGIT)+;
fragment DEL    :   SPACE ',' SPACE ; //Delimiter + may be space behind
fragment NUMBER :   (SIGN)? DIGIT_SEQ ('.' (DIGIT_SEQ)?)?; // should be given in decimal degrees, North is 0 and direction is clockwise, range is 0 to 360
fragment SIGN   :   '+' | '-';
fragment YEAR   :   DIGIT DIGIT DIGIT DIGIT;
fragment MONTH  :   DIGIT DIGIT;
fragment DAY    :   DIGIT DIGIT;
fragment HOUR   :   DIGIT DIGIT;
fragment MINUTE :   DIGIT DIGIT;
fragment SECOND :   DIGIT (DIGIT)? ('.' (DIGIT)+)?;

fragment SPACE : (' ')?;// used to increase compatibility

Thing is, I have this message when creating the AST :

line 1:11 no viable alternative at input ''%bridge%''
line 1:35 no viable alternative at input ''%Munich%''

The generated tree is though correct (as far as I’m concerned at least):

So, could anyone give me a hint about what’s wrong in there ? I think character contains all extra characters needed to correclty parse this expression. . .

Thanks !

As usual, some Java code to quickly test the grammar :

import org.antlr.runtime.*;
import org.antlr.runtime.tree.*;
import org.antlr.stringtemplate.*;

public class Main {
  public static void main(String[] args) throws Exception {

    // the expression
    String src = "(ICOM LIKE '%bridge%' or ICOM LIKE '%Munich%')";

    // create a lexer & parser
    //LogicLexer lexer = new LogicLexer(new ANTLRStringStream(src));
    //LogicParser parser = new LogicParser(new CommonTokenStream(lexer));

    LogicLexer lexer = new LogicLexer(new ANTLRStringStream(src));
    LogicParser parser = new LogicParser(new CommonTokenStream(lexer));

    // invoke the entry point of the parser (the parse() method) and get the AST
    CommonTree tree = (CommonTree)parser.parse().getTree();

    // print the DOT representation of the AST 
    DOTTreeGenerator gen = new DOTTreeGenerator();
    StringTemplate st = gen.toDOT(tree);
    System.out.println(st);
  }
}

Answer 1

I see 3 issues:

1

Your atom rule matches epsilon (nothing):

atom   
 : ID 
 | | '(' expression ')' -> expression
 ;

(note the “nothingness” inside | |)

causing your grammar to be ambiguous. I guess it should be:

atom   
 : ID 
 | '(' expression ')' -> expression
 ;

2

Your fragment CHARACTER matches a single quote while this single quote also denotes the end of the fragment QUOTED_STRING.

I guess CHARACTER should be this instead:

fragment CHARACTER : ('a'..'z' | 'A'..'Z' | '.' | '%'); 

3

Nowhere in your parser rule you match the token CONSTANT_EXPRESSION, so the AST you posted could never have been created by a parser generated from the grammar you posted. I presume you’d want to match it in the atom rule like this:

atom   
 : ID 
 | CONSTANT_EXPRESSION
 | '(' expression ')' -> expression
 ;

With the changes outlined above, I get the following AST without any errors being printed to the console: