Adam Hinchliffe wrote:

>  Hi,
>  
>  I am new to ruby and am trying to scrape a website table into a matrix,
>  I have been playing around with WWW::Mechanize and have had some success
>  getting the page, extracting the table I want and then separating the
>  result by table rows.  The problem comes with then splitting it down by
>  table data, I am going through each result in the array to break it by
>  the <td> tag, but my code appears to have zero effect!
>  
>  The code I am using is below; any help in getting the table into a
>  matrix would be really appreciated.

/ ... snip code listing

Try this:

-----------------------------------------
#!/usr/bin/ruby -w

table = "<table> <tr>\n" +
"<td> 4</td><td>47</td><td>1</td><td>19</td></tr>\n" +
"<tr> <td>7</td><td>49</td><td>4</td><td>39</td></tr>\n" +
"<tr> <td>14</td><td>17</td><td>19</td><td>21</td>\n" +
"</tr> </table>\n"

rows = table.scan(%r{<tr> .*?</tr>}m)

rows.each do |row|
   fields = row.scan(%r{<td> (.*?)</td>}m)
   puts fields.join(",")
end
-----------------------------------------

Output:

4,47,1,19
7,49,4,39
14,17,19,21

Try this filter on a more complex page, one with differing numbers of cells
in each row, etc. It works quite well, and you can understand what it is
doing at a glance.

And notice that most of the listing is sample data, the actual filter code
consists of five lines of Ruby.

The important thing to remember about Ruby is that actually writing code is
so efficient that it is hard to justify applying a predefined package to
some of the simpler processing tasks.

-- 
Paul Lutus
http://www.arachnoid.com