Table.LoadXML and column width setting

[mjcoon]

I have a program that shows data in a table and uses two views of the table with most of the columns visible (one is not meant for viewing) or a subset.

I allow the data to be saved and restored via CSV or XML formats.

When restoring from XML I find that all the column widths which were set by AddCol() or subsequently by ColWidth() = are now set to some default value (so that even the column meant to be invisible can be seen).

Perhaps this is just a side effect of LoadXML (though I cannot see why this should not be a bug). So I expected to get round it just by using the code that already works in reverting from subset to full viewing. But I find that after LoadXML the ColWidth() = code no longer works at all!

I attach a zipped source with two example data files in CSV and XML format. You will see that it works fine under the desktop IDE with either CSV or XML. But when compiled the fault appears.

Mike.

 

[agraham]

Succinctly the problem is, when compiled, once a table has been loaded by LoadXml none of the columns can have their width set programatically even when the table is subsequently reloaded by LoadCSV. In the IDE there is no problem. All this is on the desktop.

I see this behaviour consistently but am at a loss to explain it (which probably means it is something simple but I don't really understand the .NET data controls!). Poking round with Reflector I can see no significant differences between the IDE code and the compiled code.

 

[Erel]

This is a problem/bug with the XML loading and the ability to later change the layout.
Frankly I don't see any reason for using the XML format rather than the CSV format for saving the table data (which is smaller and much faster to load).

The XML format is only kept for backwards compatibility.

 

[mjcoon]

This is a problem/bug with the XML loading and the ability to later change the layout.
Frankly I don't see any reason for using the XML format rather than the CSV format for saving the table data (which is smaller and much faster to load).

The XML format is only kept for backwards compatibility.

Hi Erel, that's fair enough, I have no real need to use XML at all.

Is there any possibility of having column heading texts separate from the column names used in the code?

 

[Erel]

Is there any possibility of having column heading texts separate from the column names used in the code?

That is only possible with SQL tables, where you can set the name of each column in the query.

 

[agraham]

This is a problem/bug with the XML loading and the ability to later change the layout.

What I don't understand is why almost identical code works in the IDE but not when compiled. I can see no real difference but maybe I'm overlooking something.

 

[Erel]

I actually don't really know the exact cause. The way it works with the DataSet and DataTable is far too complex (for me at least).