Question

I need to generate a report in csv format. Data is retrieved as List<T> and only certain columns needs to be displayed in the report. I am thinking of doing it by storing its schema in an XML document under App_Data folder and use LINQ to XML to retrieve the field names and create the report.

Sample XML:

<report>
<fields>
<field headercaption='Customer Name'>CName</field>
<field headercaption='Address'>CAddress</field>
...
</fields>
</report>

Is it advisable to completely depend on XML file this way or do I need to do it through coding.

Edit

Fields are properties of List<T> which needs to be populated in the CSV files. Header caption is the name of the column for the field in the report.

Sample Report

Customer Name, Address,

ALFKI , 31 Independence Ave., Washington

Answer 1

If you just want to add/remove columns at run time in this one report, it would be even easier to use a plain text file. Store column information in a format you can easily parse ( something like CName|Customer Name). When you need to generate CSV, read all lines, split by | (or whatever separator you choose) and go from there. You will probably also want to add a comment or documentation note somewhere about the format (in case you or someone else need to come back and modify it in a year or more).

XML might become useful if you:

• plan to put several reports into one file, and maybe give them names
• provide some way for other people to submit their own reports
• create some authoring tool to allow people to customize reports
• localize column names in CSV. You could define something like

.

<report>
<fields>
   <field name="CName">
      <header language="en-us">Customer Name</header>
      <header language="es-es">Nombre de cliente</header>
      ...
   </field>
   ...
</fields>
</report>

• want to experiment with XML libraries

In those cases XML may be useful because you could define an XSD and easily validate that submitted reports are valid.