Navigating through the HTML table using code

[Sridhar Venkataraman]

Say I have a HTML table rendered onto a web page with a certain number of rows and columns. Is it possible to "navigate" through the table (i.e. to particular row-column intersections [cells]) using code? Can someone please give me a sample?

Thanks a ton!

[Rob Spoor]

The javax.swing.text.html and javax.swing.text.html.parser packages can help you out a bit.

Basically, you'll need a parser, and a callback to handle what you encounter:

Reader reader; // initialize with the contents
HTMLEditorKit.Parser parser = new ParserDelegator();
HTMLEditorKit.ParserCallback cb = new HTMLEditorKit.ParserCallback()
{
    public void handleStartTag(HTML.Tag tag, MutableAttributeSet atts, int pos)
    {
        if (HTML.Tag.TABLE.equals(tag))
        {
            // you're now entering a table element
        }
    }
};
parser.parse(reader, cb, true);

[Stan James]

This may get thrown out of the beginner forum, but for a real mind blower, see the Parse class in the FIT testing tool. It parses an HTML table into a graph of objects in only a few lines.

[Sridhar Venkataraman]

I had no idea how difficult this is to do with Java. I have to do this in a web application using JSP.

What is the FIT testing tool? I am totally uninitiated!

[Bear Bibeault]

So I take it you are doing some screen scraping? Cannot you negotiate with your data provider to get the information in a more digestible format? HTML is a display format so it's not surprising that it's difficult to try and use it as a data format.

[Bert Bates]

as predicted, this is SO getting thrown out of the beginner's forum

[Sridhar Venkataraman]

Thanks for the move and the replies..

I actually have asked this question because it is the only idea I got in my head in relation to solving the *big* dilemma I posted here. Any help there will be appreciated whole-heartedly!

https://coderanch.com/t/292896/JSP/java/Seating-arrangement-application-JSP-Struts

[Stan James]

I followed up in the other thread. I think you need to generate HTML, not parse it. Does that sound right?