PowerShell - xpath for Case Insensitive XML attribute name

Question

I have a XML where the attribute names can be combinations of lower-case and upper-case letters. In the below example, 'datasource' attribute can have any number of lowercase and uppercase letters.

I need to fetch those nodes where the 'datasource' is XML. I have searched all over the internet, but could not find any solution for this. Several examples are there for translate(), lower-case() but they don't fit my scenario.

[xml] $GM_ProcessXML =@'
<Process>
    <Parameter Name="Parameter1" Datasource="XML"><![CDATA[Sujeet]]></Parameter>
    <Parameter Name="Parameter2" DataSource="XML"><![CDATA[Padhi]]></Parameter>     
    <Parameter Name="Parameter3" DatASource="XML"><![CDATA[Padhi]]></Parameter>     
    <Parameter Name="Parameter4" datASource="XML"><![CDATA[Padhi]]></Parameter>     
    <Node>
        <Node1 Name="Node1" Datasource="XML"><![CDATA[Sujeet]]></Node1>
        <Node2 Name="Node2" DataSource="XML"><![CDATA[Padhi]]></Node2>      
        <Node3 Name="Node3" DatASource="XML"><![CDATA[Padhi]]></Node3>      
        <Node4 Name="Node4" datASource="XML"><![CDATA[Padhi]]></Node4>    
    </Node>
</Process>
'@

$XPath = "//*[@datasource='XML']"

$Nodes = $GM_ProcessXML.SelectNodes($XPath)

$Nodes

In xml 2.0 you can do this: "//*[lower-case(@datasource)='xml']" — js2010, Jun 05 '20 at 13:45

score 1 · Answer 1 · answered Jun 05 '20 at 11:22

1

You could also use the case-insensitive dot notation like this:

$Nodes = $GM_ProcessXML.Process.Parameter | Where-Object { $_.DAtaSourcE -eq 'xML' }

This will return an array of System.Xml.XmlElement nodes

answered Jun 05 '20 at 11:22

Theo

57,719
8
24
41

Actually the XML is far more complex and I need all the Nodes where the Datasource is XML. So I am looking for a xpath. Edited the XML. @Theo – Sujeet Padhi Jun 05 '20 at 12:28

score 0 · Answer 2 · edited Jun 05 '20 at 15:01

Short of modifying the files:

[xml]$GM_ProcessXML = @'
<Process>
    <Parameter Name="Parameter1" Datasource="XML"><![CDATA[Sujeet]]></Parameter>
    <Parameter Name="Parameter2" DataSource="XML"><![CDATA[Padhi]]></Parameter>
    <Parameter Name="Parameter3" DatASource="XML"><![CDATA[Padhi]]></Parameter>
    <Parameter Name="Parameter4" datASource="XML"><![CDATA[Padhi]]></Parameter>
    <Node>
        <Node1 Name="Node1" Datasource="XML"><![CDATA[Sujeet]]></Node1>
        <Node2 Name="Node2" DataSource="XML"><![CDATA[Padhi]]></Node2>
        <Node3 Name="Node3" DatASource="XML"><![CDATA[Padhi]]></Node3>
        <Node4 Name="Node4" datASource="XML"><![CDATA[Padhi]]></Node4>
    </Node>
</Process>
'@ -replace 'datasource','datasource'

$XPath = "//*[@datasource='XML']"

$Nodes = $GM_ProcessXML.SelectNodes($XPath)

$Nodes


Name       datasource #cdata-section
----       ---------- --------------
Parameter1 XML        Sujeet
Parameter2 XML        Padhi
Parameter3 XML        Padhi
Parameter4 XML        Padhi
Node1      XML        Sujeet
Node2      XML        Padhi
Node3      XML        Padhi
Node4      XML        Padhi

Sorry for the edit, that was completely unintended. (I thought I was editing my own answer, I wasn't paying attention) — Tomalak, Jun 05 '20 at 15:01
Of course, running search-and-replace operations on XML is dangerous (and wrong) an should be avoided at all costs. — Tomalak, Jun 05 '20 at 15:04

score 0 · Accepted Answer · answered Jun 05 '20 at 14:53

There is no case-insensitive string comparison in XPath 1.0
XPath 1.0 is all that is available in .NET
PowerShell is based completely on .NET

So, on the face of it, you are out of luck. But there is a work-around by combining PowerShell code and XPath translate().

[xml] $GM_ProcessXML =@'
<Process>
    <Parameter Name="Parameter1" Datasource="XML"><![CDATA[Sujeet]]></Parameter>
    <Parameter Name="Parameter2" DataSource="XML"><![CDATA[Padhi]]></Parameter>     
    <Parameter Name="Parameter3" DatASource="XML"><![CDATA[Padhi]]></Parameter>     
    <Parameter Name="Parameter4" datASource="XML"><![CDATA[Padhi]]></Parameter>     
    <Node>
        <Node1 Name="Node1" Datasource="XML"><![CDATA[Sujeet]]></Node1>
        <Node2 Name="Node2" DataSource="XML"><![CDATA[Padhi]]></Node2>      
        <Node3 Name="Node3" DatASource="XML"><![CDATA[Padhi]]></Node3>      
        <Node4 Name="Node4" datASource="XML"><![CDATA[Padhi]]></Node4>    
    </Node>
</Process>
'@

$attributeName = 'DataSource'
$lc = $attributeName.ToLowerInvariant()    # -> 'datasource'
$uc = $attributeName.ToUpperInvariant()    # -> 'DATASOURCE'

$XPath = "//*[@*[translate(name(), '$uc', '$lc') = '$lc'] = 'XML']"
# ->      //*[@*[translate(name(), 'DATASOURCE', 'datasource') = 'datasource'] = 'XML']

$Nodes = $GM_ProcessXML.SelectNodes($XPath)

Of course you can write the translate function by hand, especially in cases where the string to translate is basically hard-coded.

But this approach is generic, so you can also use it to search case-insensitively for any user-provided/dynamic value, e.g.:

$value = 'sujeet'   # as the user has entered it

$lc = $value.ToLowerInvariant()
$uc = $value.ToUpperInvariant()

$XPath = "//Parameter[translate(., '$uc', '$lc') = '$lc']"
# ->      //Parameter[translate(., 'SUJEET', 'sujeet') = 'sujeet']

This is a lot better than the often-proposed

translate(., 'ABCDEFGHIJKLMNOPQRSTUVWXYZ', 'abcdefghijklmnopqrstuvwxyz')

since that is limited to pre-defined alphabets, and fails for e.g. accented characters unless you explicitly include them, which quickly becomes unwieldy.

PowerShell - xpath for Case Insensitive XML attribute name

3 Answers3