Reading XML files with child elements
In the previous post we looked at how to create a DataSet by reading from a simple XML file. In this port we will investigate how .NET handles more complex XML files and in particular how the .NET DataSet handles XML files containing multiple child elements under the root.
<?xml version="1.0" encoding="utf-8" ?>
<AppSettings>
<setting name="CurrentInstance" serializeAs="String">
<value>1</value>
<currentStatus>
<status>live</status>
<lastUpdated>01-01-2010</lastUpdated>
</currentStatus>
</setting>
<setting name="OutputDirectory" serializeAs="String">
<value>\\SV-OCRMGR\mbrc\ocr\xmloutput</value>
<currentStatus>
<status>live</status>
<lastUpdated>01-01-2010</lastUpdated>
</currentStatus>
</setting>
</AppSettings>
As we found in the previous post, when .NET parses the XML into the DataSet it creates a DataTable for the root element, in thise case the table is called AppSettings. With the more complex XML shown above, simple elements, like value are parsed into a column of the AppSettings table. Elements that contain child elements themselves, such as currentStatus are parsed into a new DataTable within the DataSet. .NET also creates a relationship, similar to a foreign key relationship you may be familiar with from database development.
We can modify the code from the previous post to take advantage of this functionality.
//declare a new DataSet
DataSet ds = new DataSet();
//read the XML file into the DataSet
ds.ReadXml("AppSettings.xml");
DataRow[] drChildren;
DataRelation dr;
//access each node in the XML file using a DataRow
//in this example the nodes we want are in the settings table
foreach (DataRow drResult in ds.Tables["setting"].Rows)
{
//access the elements of the XML file using the DataRow columns
//we will just write them to Trace for now
System.Diagnostics.Trace.Write(drResult["name"].ToString() + " " + drResult["value"].ToString());
//declare the relationship
dr = ds.Relations[0];
//although we only expect one row of data GetChildRows returns an array
drChildren = drResult.GetChildRows(dr);
//iterate thru the array of DataRows to access the values
for (int i = 0; i < drChildren.Length; i++)
{
System.Diagnostics.Trace.Write(" " + drChildren[i]["status"].ToString());
}
System.Diagnostics.Trace.WriteLine("");
}
The output from this file (show below) shows that we have returned the status element for each row in the setting table:
CurrentInstance 1 live
OutputDirectory \\SV-OCRMGR\mbrc\ocr\xmloutput live
What isn't clear is how all this works. To illustrate this better, we can write some code to reveal the relationship .NET created between our original setting table and the child element currentStatus.
//declare a new DataSet
DataSet ds = new DataSet();
//read the XML file into the DataSet
ds.ReadXml("AppSettings.xml");
//iterate through the relationships
foreach (DataRelation dr in ds.Relations)
{
System.Diagnostics.Trace.WriteLine(dr.RelationName + " " + dr.ParentTable + " " + dr.ChildTable);
foreach (DataColumn dc in dr.ParentColumns)
{
System.Diagnostics.Trace.WriteLine(dc.ColumnName);
}
}
The output of this codes is:
setting_currentStatus setting currentStatus
setting_Id
.NET has created a DataRelation called setting_currentStatus which links the setting table to the currentStatus table in a field called setting_Id. You'll notice the XML file does not have a field called setting_Id, .NET created that when it parsed the file into the DataSet. In our earlier code we took advantage of this relationship by calling the GetChildRows method on each DataRow in setting. GetChildRows takes as its parameter a DataRelation object, which tells .NET everything it needs to know to grab the related child rows.
As we only had one complex child element in our XML, we were able to use the define the correct DataRelation by simply pointing to the first DataRelation in the collection using the code dr = ds.Relations[0]. Obvisously, when our XML has more than one complex child element, we need a way to determine the DataRelation we want to use. Lets modofy the XML file again and add a new element called currentScope.
<?xml version="1.0" encoding="utf-8" ?>
<AppSettings>
<setting name="CurrentInstance" serializeAs="String">
<value>1</value>
<currentStatus>
<status>live</status>
<lastUpdated>01-01-2010</lastUpdated>
</currentStatus>
<currentScope>
<scope>user</scope>
<lastUpdated>01-01-2010</lastUpdated>
</currentScope>
</setting>
<setting name="OutputDirectory" serializeAs="String">
<value>\\SV-OCRMGR\mbrc\ocr\xmloutput</value>
<currentStatus>
<status>live</status>
<lastUpdated>01-01-2010</lastUpdated>
</currentStatus>
<currentScope>
<scope>application</scope>
<lastUpdated>01-01-2010</lastUpdated>
</currentScope>
</setting>
</AppSettings>
We have a couple of options when defining the DataRelation, one is to use the index as we did previously, which will return the related children in the order they appear in the XML. The second is to use the name of the DataRelation, which as we saw earlier is created by combining the names of the parent and child tables. Either way, we need to know the structure of the XML file and be confident that it is not going to change too dramatically. Lets modify the code to use the DataRelation name to retrieve the child rows from the newly added currentScope element.
//declare a new DataSet
DataSet ds = new DataSet();
//read the XML file into the DataSet
ds.ReadXml("AppSettings.xml");
DataRow[] drChildren;
DataRelation dr;
//access each node in the XML file using a DataRow
//in this example the nodes we want are in the settings table
foreach (DataRow drResult in ds.Tables["setting"].Rows)
{
//access the elements of the XML file using the DataRow columns
//we will just write them to Trace for now
System.Diagnostics.Trace.Write(drResult["name"].ToString() + " " + drResult["value"].ToString());
//declare the relationship
dr = ds.Relations["setting_currentScope"];
//although we only expect one row of data GetChildRows returns an array
drChildren = drResult.GetChildRows(dr);
//iterate thru the array of DataRows to access the values
for (int i = 0; i < drChildren.Length; i++)
{
System.Diagnostics.Trace.Write(" " + drChildren[i]["scope"].ToString());
}
System.Diagnostics.Trace.WriteLine("");
}
The output from this code:
CurrentInstance 1 user
OutputDirectory \\SV-OCRMGR\mbrc\ocr\xmloutput application