The default format used for the
DataSet XML works
perfectly well when you create a new cross-platform application.
However, most systems include legacy clients that require data in a
set format. Unfortunately, even a minute difference between the
expected and the actual XML file format can prevent a client from
successfully reading the data.
ADO.NET offers limited ways to customize the generated XML for a
DataSet. Before writing your own custom code, you
should begin by examining these features. If you need to perform more
significant changes, you need to perform these additional operations
using the .NET XML classes or XLST transformation.
One recurring question in XML modeling is the question of whether to use attributes to store data or contained elements. For example, here’s a category row that uses only elements:
<Categories> <CategoryID>1</CategoryID> <CategoryName>Beverages</CategoryName> <Description>Soft drinks, coffees, teas, beers, and ales</Description> </Categories>
And here’s the same row with attributes:
<Categories CategoryID="1" CategoryName="Beverages" Description="Soft drinks, coffees, teas, beers, and ales" />
With ADO.NET, you can configure whether column values are stored in
attributes or elements. Best of all, you can do it on a per-column
basis (for example, storing the unique identity column as an
attribute and all other columns as elements). All you need to do is
DataColumn.ColumnMapping property to one of the ...