I agree with Dan 100%.
I'd also offer an alternative solution. Before Tablesorter 2.0, I had
a similar situation and resolved it by adding a hidden <span> tag
before the data that contained the 'raw' timestamp integer for sorting
purposes. This worked really well with the old tablesorter (and still
works with 2.0).
Here's what I have:
<td><span style="display: hidden;">12345</span>1 day 2 hours ......</
td>
The table sorter parses the first bit of "data" it sees.
On Jul 8, 6:51 am, "Dan G. Switzer, II" <[EMAIL PROTECTED]>
wrote:
> >I am using the tablesorter plugin, which I really like. However, in
> >several of my table cells I have oddly formatted time values that look
> >like the following:
>
> >1d 12h 34m 25s
>
> >Which maps to:
>
> >1 day, 12 hours, 34 mins and 25 seconds
>
> >I cannot change how these values are displayed - they have to be in
> >that format. However, in my code I also create an attribute to the
> >table cell called sorttable_customval in which I insert the raw number
> >format as a float:
>
> ><td sorttable_customkey="dlt" sorttable_customval="65.093" >1m 5s</
> >td>
>
> > I have tried to write my own add-on to the tablesorter source code to
> >get this attribute value in the confines of the tablesorter
> >textExtraction method:
>
> >Javascript at the top of my webpage with the table
> >======================================
> >("#monitor").tablesorter({
> > textExtraction: 'customattribute',
> > debug: true,
> > widgets:['zebra']
> >})
>
> >The modified part of tablesorter.js (rows 233 and 234)
> >=========================================
> >221 function getElementText(config,node) {
> >222
> >223 if(!node) return "";
> >224
> >225 var t = "";
> >226
> >227 if(config.textExtraction == "simple") {
> >228 if(node.childNodes[0] &&
> >node.childNodes[0].hasChildNodes()) {
> >229 t = node.childNodes[0].innerHTML;
> >230 } else {
> >231 t = node.innerHTML;
> >232 }
> >233 } else if(config.textExtraction ==
> >"customattribute") {
> >234 t = $(node).attr('sorttable_customval');
> >235 } else {
> >236 if(typeof(config.textExtraction) ==
> >"function") {
> >237 t = config.textExtraction(node);
> >238 } else {
> >239 t = $(node).text();
> >240 }
> >241 }
> >242 return t;
> >243 }
>
> >I thought this would work, but I get the following error in Firebug:
>
> >s is undefined jquery.tablesorter.js (line 703)
> > return $.trim(s.toLowerCase());
>
> The problem with using the textExtraction (at least from my understanding)
> is that it works for all columns. So, I suspect the error is coming from the
> fact that the 'sorttable_customval' only existing on some columns and is
> returning undefined for the other columns (thus breaking $.trim().)
>
> You were close when you wrote a custom parser--which is the way I've solved
> this problem in the past. The problem is you used the "is" option instead of
> the "function" option.
>
> Try this:
>
> $.tablesorter.addParser({
> id: 'dlt',
> is: function (s){ return false },
> format: function(value, table, cell){
> return cell.parentNode.getAttribute('sorttable_customval');
> },
> type: 'numeric'
>
> });
>
> In the "format" option, the "value" is what's coming back from the
> textExtraction method. In this case, we want to ignore this value all
> together, and instead use the cell to extract the attribute.
>
> -Dan
