Thread: java parser

Post your code and ask specific questions. Noboidy will provide ready made code.

The String class has some useful methods for separating parts of a String into tokens. What were your problems with its methods?
Also the StreamTokenizer has some useful methods for doing the same.
The semantics of the tokens is up to the programmer.

Start by defining the syntax of the statements you want to parse so the tokens can be organized.

Does the position of these verbs and nouns change in the sentence?

Have you looked at defining a syntax for the statements that you are trying to parse?

You cannot decide what to use until you have finalized the syntax. You haven't answered Norm's question.

It all depends on the syntax and how the parsing is to be done. For example are all the tokens delimited by spaces?
The split() method uses a regular expression that can be powerful for some parsing. The StringTokenizer can return delimiters which can be useful. The StreamTokenizer has several useful methods that can tell you about the next token it has found.

Writing a parser for a statement is easier if that statement's contents are well defined.
Consider the java language. It has several different types of statements that are all well defined. The parser uses the syntax rules as it reads the tokens in a statement to make decisions about where each token fits to make a valid statement.
So far the statements you have posted all start with: "There"
followed by either "is" or "are"
and so on through the rest of the statement.

Writing a parser is not a first year programming project.

find a method to generalize the input you parse. What is the general structure you're parsing? what is easy (the number of items, location)? what is difficult (dimensions)?

EDIT: Norm's reply is better. Read that.

I haven't worked with writing parsers for years. In school we studied different language syntax rules and used BNF to define a language for a compiler writing project. I haven't done much with those techniques in a long time.

 import java.util.StringTokenizer;
 
public class Parser {
	public static void main(String args[]){
 
		parser("There is a table (dimension: 4 in 5 in 3 meters) in the room.");
 
 
	}
 
	public static void parser(String str){
 
		String type = new String("Type: ");
		String number = new String("Number: ");
		String size = new String("Size: ");
		String position = new String("Position: ");
 
		StringTokenizer strTok = new StringTokenizer(str, " ");
		String[] wordList = new String[strTok.countTokens()];
 
 
			if(str.substring(6, 8).equals("is")){
				System.out.println(number + "1");
			}
 
		}
	}

Here I first declared the type, number, size and position. Then I used StringTokenizer to split the string based on spaces. Then I counted the number of the tokens ( I'm not sure that this works properly ) Then I tell that If the verb was "is" print out: Number: 1.
Now I want to say that if the verb was "are" please print out the third token as number. How can I say that? Can I put all tokens in an Arraylist and then tell that print the third array? Or is it a good opinion at all?

Please edit your post and wrap your code with code tags:
[code=java]
YOUR CODE HERE
[/code]
to get highlighting and preserve formatting.

I want to say that if the verb was "is" please print out the third token as number.

I'm not sure why you are counting the tokens. If the current token is: "is" then the next token needs to be "a"
and the number of items would be set to 1.
The parser walks through the statement token by token. The contents of a token will define what needs to come next. "is" is followed by "a', "are" is followed by a number.

if(str.substring(6, 8).equals("is")){

That statement makes no sense. "is" will be the whole token, not part of a larger String.

I don't know what use counting the tokens will be to help parse the statement.

If the current token is "are" then the next token must be a number.
Notice I used current and next to describe the tokens, not second and third.
(I'm assuming that is the syntax of a valid statement.)

The syntax for a valid statement should show the relative position of the tokens, not their numeric location.

But it doesn't work

Please explain.

What is strTok? Is it a String or is it an instance of a method that has a nextToken() method?