Getting html table cell values based on the table head

Hi, 

 

Firstly I would like to do this as an xpath expression, if this is an
inappropriate question for this forum then I apologise, but would like to
see an XSL equivalent also.  

 

The data is like this...

 

<table border="1" width="500">

  <tbody>

    <tr>

      <th class="awrbg">DB Name</th>

      <th class="awrbg">DB Id</th>

      <th class="awrbg">Instance</th>

      <th class="awrbg">Inst num</th>

      <th class="awrbg">Release</th>

      <th class="awrbg">RAC</th>

      <th class="awrbg">Host</th>

    </tr>

 

    <tr>

      <td class="awrnc">SFDB</td>

      <td class="awrnc" align="right">3911180340</td>

      <td class="awrnc">atlsfdb</td>

      <td class="awrnc" align="right">1</td>

      <td class="awrnc">10.2.0.5.0</td>

      <td class="awrnc">NO</td>

      <td class="awrnc">sfd02</td>

    </tr>

  </tbody>

</table>

 

I would like to be able to construct a single xpath expression which
supplies the "index" e.g. "DB Name" which is on path
/table/tbody/tr[1]/th[1] and returns the value from the corresponding nth
row in this case the second tr of the same /table/tbody so the value
"SFDB". 

 

Input of "Host" returns "sfd02" .  etc.

 

 

Then I would like to extend this to cover tables with any number of rows
(and an inconsistency in the use of "th" and "tr" for headings) e.g. 

 

 

        <table border="1">

            <tbody>

                <tr>

                    <th class="awrbg">Wait Class</th>

                    <th class="awrbg">Waits</th>

                    <th class="awrbg">%Time -outs</th>

                    <th class="awrbg">Total Wait Time (s)</th>

                    <th class="awrbg">Avg wait (ms)</th>

                    <th class="awrbg">%Total Call Time</th>

                </tr>

 

                <tr>

                    <td class="awrc">Application</td>

                    <td class="awrc" align="right">625,444</td>

                    <td class="awrc" align="right"/>

                    <td class="awrc" align="right">31,580</td>

                    <td class="awrc" align="right">50</td>

                    <td class="awrc" align="right">53.03</td>

                </tr>

 

                <tr>

                    <td class="awrnc">CPU time</td>

                    <td class="awrnc" align="right"/>

                    <td class="awrnc" align="right"/>

                    <td class="awrnc" align="right">27,737</td>

                    <td class="awrnc" align="right"/>

                    <td class="awrnc" align="right">46.58</td>

                </tr>

 

                <tr>

                    <td class="awrc">Network</td>

                    <td class="awrc" align="right">34,576,138</td>

                    <td class="awrc" align="right"/>

                    <td class="awrc" align="right">2,565</td>

                    <td class="awrc" align="right">0</td>

                    <td class="awrc" align="right">4.31</td>

                </tr>

 

                <tr>

                    <td class="awrnc">System I/O</td>

                    <td class="awrnc" align="right">6,850,776</td>

                    <td class="awrnc" align="right"/>

                    <td class="awrnc" align="right">636</td>

                    <td class="awrnc" align="right">0</td>

                    <td class="awrnc" align="right">1.07</td>

                </tr>

 

                <tr>

                    <td class="awrc">Commit</td>

                    <td class="awrc" align="right">413,851</td>

                    <td class="awrc" align="right"/>

                    <td class="awrc" align="right">472</td>

                    <td class="awrc" align="right">1</td>

                    <td class="awrc" align="right">0.79</td>

                </tr>

 

                <tr>

                    <td class="awrnc">User I/O</td>

                    <td class="awrnc" align="right">1,012,139</td>

                    <td class="awrnc" align="right"/>

                    <td class="awrnc" align="right">323</td>

                    <td class="awrnc" align="right">0</td>

                    <td class="awrnc" align="right">0.54</td>

                </tr>

 

                <tr>

                    <td class="awrc">Other</td>

                    <td class="awrc" align="right">13,976</td>

                    <td class="awrc" align="right">1</td>

                    <td class="awrc" align="right">53</td>

                    <td class="awrc" align="right">4</td>

                    <td class="awrc" align="right">0.09</td>

                </tr>

 

                <tr>

                    <td class="awrnc">Concurrency</td>

                    <td class="awrnc" align="right">55,797</td>

                    <td class="awrnc" align="right"/>

                    <td class="awrnc" align="right">13</td>

                    <td class="awrnc" align="right">0</td>

                    <td class="awrnc" align="right">0.02</td>

                </tr>

 

                <tr>

                    <td class="awrc">Configuration</td>

                    <td class="awrc" align="right">63</td>

                    <td class="awrc" align="right">5</td>

                    <td class="awrc" align="right">1</td>

                    <td class="awrc" align="right">12</td>

                    <td class="awrc" align="right">0.00</td>

                </tr>

 

                <tr>

                    <td class="awrnc">Administrative</td>

                    <td class="awrnc" align="right">4</td>

                    <td class="awrnc" align="right"/>

                    <td class="awrnc" align="right">0</td>

                    <td class="awrnc" align="right">9</td>

                    <td class="awrnc" align="right">0.00</td>

                </tr>

            </tbody>

        </table>

 

So I would like to be able to provide an input of "Wait Class" and either 

-          (if document order can be relied on) get an ordered set of
"Application", "CPU time", "Network", "System I/O" , "Commit", "User I/O",
"Other", "Concurrency", "Configuration", "Administrative"

-          Or supply an "index" value of 1 and get back "Application" and
then supply an index of 2 and get back "CPU time" 

(I am assuming that I can rely on the first option across all
implementations?).

 

In case anyone is wondering - these are Oracle AWR reports which I am
trying to parse, I have thousands of historical ones and need to build
some trend analysis based on them. 

 

Cheers, 

Karl.