Dojox Datagrid, add span to every cell to make css nowrap to work in IE7 - dojox.grid.datagrid

Question

Since nowrap on td element doesn't seems to work in IE, see this question, I am in desperate need of help how to add a span element with nowrap to every cell in a Dojox Datagrid without having to define field formatters to accomplish this.
See jsfiddle here http://jsfiddle.net/HkxHZ/4/
Using this css I get what I want in Chrome and FF, i.e. no word wrap and overflow hidden. But it doesn't work in IE..
<style type="text/css">
.dojoxGridRowTable {
table-layout: fixed;
width: 0px;}
.dojoxGrid .dojoxGridCell {
text-align: left;
overflow:hidden;
white-space:nowrap;}
</style>​

Answer 1

This method is like auto-size in that is will make the width of each column in the DataGrid big enough to show your data with no wrapping.
What you should do is calculate the proper width for each column of the grid layout based on PX size of the text in the column description and data for that column. This should be done before creating the grid layout.
Here is how you get the proper width in px. Add the following html to your page:
css:
#test
{
position: absolute;
visibility: hidden;
height: auto;
width: auto;
}
<div id="test"></div>
1) Get the font size used in your DataGrid
2) Get the div object and set the font size to what is used in the grid
<div id="test"></div>
var tst_item = window.document.getElementById("test");
tst_item.style.fontSize = grid_fnt_sz;
3) Place each column description and data item intended for your DataGrid into the hidden div:
tst_item.innerHTML = "Your data";
var widthPX = (tst_item.clientWidth + 1); //Width of text in PX
Store this widthPX in an array for each column keeping only the largest found for the column.
4) When creating the layout for your grid, set the width for the column to the largest width from your column description and data for that column.
This method ensures that width of each column will be big enough to show your data without wrapping. Depending on your needs, you can tweek the logic to do what you need. This might not be feasible with large amounts of data. But, it works great for me.