使用Linq搜索Xml文件，获取与子节点匹配的节点

Question

I have a Xml file with the following structure:

<Packages>
    <Package>
        <Name>Package1</Name>
        <Dependencies>
            <Dependency>Package2 </Dependency>
        </Dependencies>
    </Package>
    <Package>
        <Name>Package2</Name>
        <Dependencies>
            <Dependency>Package3</Dependency>
        </Dependencies>
    </Package>
    <Package>
        <Name>Package3</Name>
        <Dependencies />
    </Package>
</Packages>

I want to parse the Xml file, and print out the dependency "flow".

For example:

Package1 -> 
  Package2 ->
    Package3
Package2 ->
  Package3
Package3

The purpose is to create a "sequence" that will install applications, the commands should be stored in a Xml file. A package could have dependencies.

Can I use Linq to solve this problem?

I don't need a complete solution, just a hint in right direction.

Answer 1

The simple answer is yes you can. Using XPath as well.

using System.Xml.XPath;

You could set it up using recursion like this:

XElement root = XElement.Load(file); // or .Parse(xmlString);
foreach(XElement package in root.Descendants("Package"))
{
    DoPackage(package, 0);
}

void DoPackage(XElement package, int level)
{
    StringBuilder name = new StringBuilder();
    for(int i = 0; i < level; i++) name.Append("  ");
    name.Append(package.Element("Name").Value);
    var dependencies = package.Descendants("Dependency");
    if(dependencies.Count() > 0)
        name.Append(" ->");
    Console.WriteLine(name.ToString());
    foreach(XElement dependent in dependencies)
    {
        string dependentName = dependent.Value;
        XElement dependentElement = 
            package.XPathSelectElement("//Package[Name='"+dependentName+"']");
        DoPackage(dependentElement, level + 1);
    }
}

Sorry, I can't write hints very well. I didn't test or compile this, so it may need some tweaking.

There is no catching for infinite loops. If any child dependent refers to an upline package, then there will be an infinite loop.

This is also very inefficient, it would be faster if each package remembered what its dependents were and so on. For that you would have to create a class for each package. If the file you are parsing is very big, you should go with this approach. A Dictionary<string, PackageClass> with the name for the key could help.