A lot of data these days is wrapped as XML, and up until now, handling XML data wasn't a piece of cake. PowerShell makes handling XML a lot easier. This is the first part of a little series about XML and PowerShell. We start with accessing XML documents and reading data.

Getting XML Data

Let's first look how PowerShell can lay hands on XML data. One way is to create a new and empty XML document:

$xml = New-Object XML

Next, you can use your new XML document to load XML data, either from a local file or from the Internet. Use the Load() method. Let's for example load the www.powershell.com RSS ticker:

As long as you have access to the Internet, these lines will download the RSS news ticker into $a as an XML document, and when you output $a, you see a number of properties.

Browsing XML Data

What exactly is their meaning? To find out, let's first save the downloaded RSS ticker to file using the built-in Save() method. Next, we take a look at the first lines in that XML document by reading the xml file using Get-Content and selecting only the first 3 lines using Select-Object -First:

As it turns out, the PowerShell xml document in $a returns the top XML nodes: xml, xml-stylesheet and rss. Diving into nested information inside an XML document is very easy because it works in pretty much the same way as with objects. So if you'd like to see all the nodes below the root node rss, you use the rss property:

Again, if you'd looked at the raw XML data, you would see that the RSS node has these children: version, dc, slash, wfw, itunes and channel. Most of these are attributes with pieces of information. Channel however is another parent node and has a bunch of children. To look at the children, again use this property name:

Property Name Conflict - And A Simple Workaround

Bang! You get a red error message. You cannot list the children inside the channel parent node. This is a bug, or rather, a conceptional weakness in design.

If you look at the raw XML data, you will see that the channel node contains a bunch of item nodes, and these item nodes are the ones with the interesting stuff: the feed messages. Unfortunately, the XML document object already has its own item property. The moment PowerShell tries to add the item property for the item children, you get the error message. An object cannot have two properties with the same name.

And here is the workaround: work in "blind" mode and just assume the children elements are named "item" (or whatever the error message indicates):

Now you get to the good stuff! These are the headlines you're after.

Using The PowerShell Pipeline

You just learned that you can access any RSS feed in the world by loading it into a new and empty XML document.

Since RSS feeds are XML documents and adhere to a defined format, you can always list all RSS messages by looking at the RSS.Channel.Item property. This returns all raw data contained in the individual RSS items. From here, simply use PowerShell Pipeline tricks to filter what you are looking for.

First, I'd like to filter the information. I only want to see the blog entry title and the link. Easy: use Format-Table:

Next, I only want the top 5 entries, and I do not want the link to be cut off. So I insert a Select-Object -First (remember, the Format cmdlets always have to be the last element in your pipeline), and I add a -wrap parameter to Format-Table:

Can I filter based on topic, too? Sure thing. If you are interested in Windows 7, the upcoming new Windows client, and you'd like to see only blogs about Windows 7, then add a Where-Object into your pipeline.

Inside of it, the $_ placeholder represents the actual blog entry as it is running over the pipeline, and you then can check whether the title property contains one or more keywords.

Maybe you'd like to output this as a nice report. Here is the thing to remember: Never use Format-... cmdlets when you plan to output results to sources other than the console. Use Select-Object instead. To create a simple HTML report, this is what I'd do:

This is what the resulting report looks like:

Admittedly, the HTML report does not really look very stylish, but you could change that as well:

Here is more brainfood on colorizing HTML reports: http://powershell.com/cs/blogs/tips/archive/2009/01/05/outputting-html-reports.aspx

Convert Text to  XML

To wrap up this first part of our XML series, let's finally look at two more ways to read XML data into PowerShell. At the beginning of this article, we loaded the XML via Internet into an empty XML object. From here, you can save the XML as file. Using the same Load() method, you can also load XML data from an XML file:

So the second approach is to load a file-based XML into an empty XML object. The third approach uses type conversion. You read in XML data as plain text, then convert this into the XML data type.

Remember two things here: Type conversion works by writing the type (in square brackets) in front of the data you want to convert, and second: use parenthesis around Get-Content because you do not want to convert the Get-Content cmdlet of course but rather its result.

Next Steps...

Next time, we will create our own XML documents, look at some more advanced data analysis and update and change XML data in an XML file. Make sure to check back next week! Which you now by the way could automate, too. Simply use the RSS ticker in the examples above to check for new blog entries on powershell.com!

Cheers and a great and relaxing weekend to you,
and don't forget to check out PowerShell Plus! It's great and simply one of the best ways to learn PowerShell!

-Tobias

MVP Windows PowerShell