c# - Reading multiple child nodes of xml file

Question

I have created an Xml file with example contents as follows:

<?xml version="1.0" encoding="utf-8" ?>
<Periods>
  <PeriodGroup name="HER">
    <Period>
      <PeriodName>Prehistoric</PeriodName>
      <StartDate>-500000</StartDate>
      <EndDate>43</EndDate>
    </Period>
    <Period>
      <PeriodName>Iron Age</PeriodName>
      <StartDate>-800</StartDate>
      <EndDate>43</EndDate>
    </Period>
    <Period>
      <PeriodName>Roman</PeriodName>
      <StartDate>43</StartDate>
      <EndDate>410</EndDate>
    </Period>
  </PeriodGroup>
  <PeriodGroup name="CAFG">
    <Period>
      <PeriodName>Prehistoric</PeriodName>
      <StartDate>-500000</StartDate>
      <EndDate>43</EndDate>
    </Period>
    <Period>
      <PeriodName>Roman</PeriodName>
      <StartDate>43</StartDate>
      <EndDate>410</EndDate>
    </Period>
    <Period>
      <PeriodName>Anglo-Saxon</PeriodName>
      <StartDate>410</StartDate>
      <EndDate>800</EndDate>
    </Period>   
  </PeriodGroup>
</Periods>

I need to be able to read the Period node children within a selected PeriodGroup. I guess the PeriodName could be an attribute of Period if that is more sensible.

I have looked at loads of examples but none seem to be quite right and there seems to be dozens of different methods, some using XmlReader, some XmlTextReader and some not using either. As this is my first time reading an Xml file, I thought I'd ask if anyone could give me a pointer. I've got something working just to try things out, but it feels clunky. I'm using VS2010 and c#. Also, I see a lot of people are using LINQ-Xml, so I'd appreciate the pros and cons of using this method.

string PG = "HER";
XmlDocument doc = new XmlDocument();
doc.Load(Server.MapPath("./Xml/XmlFile.xml"));
string text = string.Empty;
XmlNodeList xnl = doc.SelectNodes("/Periods/PeriodGroup");
foreach (XmlNode node in xnl)
{
    text = node.Attributes["name"].InnerText;
    if (text == PG)
    {
        XmlNodeList xnl2 = doc.SelectNodes("/Periods/PeriodGroup/Period");
        foreach (XmlNode node2 in xnl2)
        {
            text = text + "<br>" + node2["PeriodName"].InnerText;
            text = text + "<br>" + node2["StartDate"].InnerText;
            text = text + "<br>" + node2["EndDate"].InnerText;
        }
    }
    Response.Write(text);
}

Answer 1

You could use an XPath approach like so:

XmlNodeList xnl = doc.SelectNodes(string.Format("/Periods/PeriodGroup[@name='{0}']/Period", PG));

Though prefer LINQ to XML for it's readability.

This will return Period node children based on the PeriodGroup name attribute supplied, e.g. HER:

XDocument xml = XDocument.Load(HttpContext.Current.Server.MapPath(FileLoc));

var nodes = (from n in xml.Descendants("Periods")
            where n.Element("PeriodGroup").Attribute("name").Value == "HER"
            select n.Element("PeriodGroup").Descendants().Elements()).ToList();

Results:

<PeriodName>Prehistoric</PeriodName>
<StartDate>-500000</StartDate>
<EndDate>43</EndDate>

<PeriodName>Iron Age</PeriodName>
<StartDate>-800</StartDate>
<EndDate>43</EndDate>

<PeriodName>Roman</PeriodName>
<StartDate>43</StartDate>
<EndDate>410</EndDate>

The query is pretty straightforward

from n in xml.Descendants("Periods")

Will return a collection of the descendant elements for the element Periods. We then use where to filter this collection of nodes based on attribute value:

where n.Element("PeriodGroup").Attribute("name").Value == "HER"

Will then filter down the collection to PeriodGroup elements that have a name attribute with a value of HER

Finally, we select the PeriodGroup element and get it's descendant nodes

select n.Element("PeriodGroup").Descendants().Elements()

EDIT (See comments)

Since the result of this expression is just a query, we use .ToList() to enumerate the collection and return an object containing the values you need. You could also create anonymous types to store the element values for example:

var nodes = (from n in xml.Descendants("Period").
             Where(r => r.Parent.Attribute("name").Value == "HER")
             select new
             {
                  PeriodName = (string)n.Element("PeriodName").Value,
                  StartDate = (string)n.Element("StartDate").Value,
                  EndDate = (string)n.Element("EndDate").Value
             }).ToList();

//Crude demonstration of how you can reference each specific element in the result
//I would recommend using a stringbuilder here..
foreach (var n in nodes)
{
      text += "<br>" + n.PeriodName;
      text += "<br>" + n.StartDate;
      text += "<br>" + n.EndDate;
}

This is what the nodes object will look like after the query has run: