Sorting and deleting XML document elements using PowerShell

I'm trying to organize an XML document that contains driver information. Here's an example of what I'm working with:

<?xml version="1.0" encoding="utf-8"?>
<IncludeFragment xmlns:p="http://schemas.microsoft.com/someschema">>
<FFUDriver>
<Component>
<Package>
<p:PackageName>Intel.Display.Driver</PackageName>
<p:PackageFeedName>Feed</PackageFeedName>
<p:Version>10.24.0.1638</Version>
<p:Flavor>release</Flavor>
</Package>
</Component>
</FFUDriver>
<FFUDriver>
<Component>
<Package>
<p:PackageName>Intel.Audio.Driver</PackageName>
<p:PackageFeedName>Feed</PackageFeedName>
<p:Flavor>release</Flavor>
<p:Version>10.24.0.1638</Version>
<p:CabName>Intel.Audio.cab</CabName>
</Package>
</Component>
</FFUDriver>
</IncludeFragment>


I need to sort each Packages' elements in the following order:

Some of the Packages' elements are already in the proper order, some aren't, as in my example XML code. Also each Package needs to be sorted in alphabetical order based on PackageName. I'm new to working with XML in PowerShell and I can't for the life of me figure out how to accomplish this.

The other requirement is to find and delete all the <CabName> elements. I sort of figured that out. The code I have below unfortunately deletes all the child elements of a <Package> element if one of its child elements is <CabName>. I can't seem to figure out the syntax to select and delete only <CabName>.

<CabName>

<Package>

<CabName>

<CabName>

$Path = 'C:Drivers.xml'
$xml = New-Object -TypeName XML
$xml.Load($Path)

$xml.SelectNodes('//Package[CabName]') | ForEach-Object {
$_.ParentNode.RemoveChild($_)
}

$xml.Save('C:Test.xml')


UPDATE: With the help of Ansgar Wiechers, here's the finished code. I updated my example XML data to include a namespace since some of the documents I work with contain them. The below code handles namespaces. I hope this helps anyone else with a similar problem/questions!

[CmdletBinding()]
Param
(
[Parameter(Mandatory = $True, Position = 0)]
[ValidateScript({
$_ = $_ -replace '"', ""
if (-Not (Test-Path -Path $_ -PathType Leaf))
{
Throw "`n `n$_ `n `nThe specified file or path does not exist. Check the file name and path, and then try again."
}
return $True
})]
[System.String]$XMLPath,

[Parameter(Mandatory = $False, Position = 1)]
[System.String]$nsPrefix = "p",

[Parameter(Mandatory = $False, Position = 2)]
[System.String]$nsURI = "http://schemas.microsoft.com/someschema"
)


# Remove quotes from full path name, if they are present
$XMLPath = $XMLPath -replace '"', ""


$xml = New-Object -TypeName XML
$xml.Load($XMLPath)
$ns = New-Object System.Xml.XmlNamespaceManager($xml.NameTable)
$ns.AddNamespace($nsPrefix, $nsURI)


# Delete all CabName elements
$xml.SelectNodes('//p:CabName', $ns) | ForEach-Object {

$_.ParentNode.RemoveChild($_) | Out-Null
}


# Sort each Package element's child nodes based on custom order
$SortList = 'p:PackageName', 'p:PackageFeedName', 'p:Version', 'p:Flavor'

$xml.SelectNodes('//Package') | ForEach-Object {

$parent = $_

$SortList | ForEach-Object {

$child = $parent.RemoveChild($parent.SelectSingleNode("./$_", $ns))
$parent.AppendChild($child)
}
} | Out-Null


# Sort each Package element in alphabetical order based on its child node PackageName
$PackageNameList = $xml.SelectNodes('//p:PackageName', $ns) | Select-Object -Expand '#text' | Sort-Object

$xml.SelectNodes('//IncludeFragment') | ForEach-Object {

$parent = $_

$PackageNameList | ForEach-Object {

$child = $parent.RemoveChild($parent.SelectSingleNode("./FFUDriver[Component/Package/p:PackageName/text()='$_']", $ns))
$parent.AppendChild($child)
}
} | Out-Null


$XMLPath = $XMLPath -replace ".xml", "_sorted.xml"

$xml.Save($XMLPath)

Write-Host "`nSorting complete. Sorted XML document saved under $XMLPath" -ForegroundColor Green


2 Answers
2

The code you have deletes all <Package> nodes that have a child element <CabName>, not just all child elements of such nodes. That's because //Package[CabName] matches all <Package> nodes that contain a <CabName> child nodes. What you actually want to match are all <CabName> nodes that have a <Package> parent node.

<Package>

<CabName>

//Package[CabName]

<Package>

<CabName>

<CabName>

<Package>

$xml.SelectNodes('./Package/CabName') | ForEach-Object {
$_.ParentNode.RemoveChild($_) | Out-Null
}


Also, normally order of elements in an XML shouldn't matter, so sorting the elements is rather pointless. However, if you for some reason you must have child nodes in a particular order you can sort the elements by removing and appending them in the desired order.

# names of the child nodes in the desired order
$nodenames = 'PackageName', 'PackageFeedName', 'Version', 'Flavor'

$xml.SelectNodes('//Package') | ForEach-Object {
$parent = $_

$nodenames | ForEach-Object {
$child = $parent.RemoveChild($parent.SelectSingleNode("./$_"))
$parent.AppendChild($child)
}
}


If you also want the <Driver> nodes sorted by package name you first need to build a sorted list of package names:

<Driver>

$xml.SelectNodes('//PackageName') | Select-Object -Expand '#text' | Sort-Object


and then use the same technique as above to remove and append the <Driver> nodes from/to the <Drivers> node. In this case you need must use a filter pattern, though

<Driver>

<Drivers>

"./Driver[Component/Package/PackageName/text()='$_']"


<Package>

<Driver>

./Driver

<Package>

.

<Package>

$parent

<Driver>

<Package>

XML converting is not needed for this job:

$xml = @"
<?xml version="1.0" encoding="utf-8"?>
<Drivers>
<Driver>
<Component>
<Package>
<PackageName>Intel.Display.Driver</PackageName>
<PackageFeedName>Feed</PackageFeedName>
<Version>10.24.0.1638</Version>
<Flavor>release</Flavor>
</Package>
</Component>
</Driver>
<Driver>
<Component>
<Package>
<PackageName>Intel.Audio.Driver</PackageName>
<PackageFeedName>Feed</PackageFeedName>
<Flavor>release</Flavor>
<Version>10.24.0.1638</Version>
<CabName>Intel.Audio.cab</CabName>
</Package>
</Component>
</Driver>
</Drivers>
"@

$XMLSorted = [System.Text.StringBuilder]::new()

$packageName = ''
$packageFeedName = ''
$version = ''
$flavor = ''

foreach( $line in @($xml -split [Environment]::NewLine) ) {

if( $line -like '*<PackageName>*' ) {
$packageName = $line
}
elseif( $line -like '*<PackageFeedName>*' ) {
$packageFeedName = $line
}
elseif( $line -like '*<Version>*' ) {
$version = $line
}
elseif( $line -like '*<Flavor>*' ) {
$flavor = $line
}
elseif( $line -like '*<CabName>*' ) {
# nothing to do
}
elseif( $line -like '*</Package>*' ) {
[void]$XMLSorted.AppendLine( $packageName )
[void]$XMLSorted.AppendLine( $packageFeedName )
[void]$XMLSorted.AppendLine( $version )
[void]$XMLSorted.AppendLine( $flavor )
[void]$XMLSorted.AppendLine( $line )
}
else {
[void]$XMLSorted.AppendLine( $line )
}
}

#Result:
$XMLSorted.ToString()


By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.