Search code examples
stringpowershellxpathsingle-quotesselect-xml

Powershell - Single Quote in String Variable with XPath


Dear Powershell fellows,

i have a (maybe) very simple problem, but i have no idea how to solve it. I want to use a String Variable, that contains a single quote, within a XPath notation. I use the CMDlet Select-Xml. If there is no single quote within the String, Select-Xml is working completly fine. But if there is one single quote (for example in: don't) it crashes my script. Let me show you in detail.

Problem

$Name = "Dont display" ##is working completly fine

(Select-Xml -Path "C:\SomePath\Somexml.xml" -XPath "//q1:Options[q1:Name = '$Name']"  -Namespace $namespace).Node.InnerText  ##is working completly fine

$Name = "Don't display" ##crashes script

(Select-Xml -Path "C:\SomePath\Somexml.xml" -XPath "//q1:Options[q1:Name = '$Name']"  -Namespace $namespace).Node.InnerText  ##crashes script

The error output of powershell is:

 Select-Xml : '//q1:Options[q1:Name = 'Don't display']' has an invalid token.
    At line:251 char:41
    + ... ing_Name = (Select-Xml -Path "C:\SomePath\Somexml.xml" -XPath ...
    +                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
        + CategoryInfo          : NotSpecified: (:) [Select-Xml], XPathException
        + FullyQualifiedErrorId : System.Xml.XPath.XPathException,Microsoft.PowerShell.Commands.SelectXmlCommand

What i tried so far

Of course i tried different quotations, such as:

 $Name = """Don't display"""
 $Name = '"Don't display"'
 $Name = "'Don't display'"
 $Name = "'Don't display'"
 $Name = 'Don't display'
$Name = ''Don't display''

It seems like, there might be a problem with powershell quotation rules and XPath notation.

But maybe anyone of you guys have an idea on how to solve it.

Thank you very much


Solution

  • It seems there's no escape character in XPath string literals, so you can't use your string delimiter inside a literal string as it terminates the string - i.e. this:

    $name = "Don't display"
    
    # single quotes:
    # //q1:Options[q1:Name = 'Don't display']
    #                            ^ terminates string literal
    

    A quick (but naive) solution to your specific issue would be to just use double-quotes as delimiters instead:

    $name = "Don't display"
    
    # double quotes:
    # //q1:Options[q1:Name = "Don't display"]
    #                            ^ *doesn't* terminate string literal
    

    but what if your data contains double quotes? Then you're back at square one..

    $name = "Don""t display"
    
    # double quotes:
    # //q1:Options[q1:Name = "Don"t display"]
    #                            ^ terminates string literal again
    

    And in a pathological case, if your literal contains both single and double quotes then you can't use either as delimiters:

    $name = "Don't d""isplay"
    
    # single quotes:
    # //q1:Options[q1:Name = 'Don't d"isplay']
    #                            ^ terminates string literal
    
    # double quotes:
    # //q1:Options[q1:Name = "Don't d"isplay"]
    #                                ^ also terminates string literal
    

    In that case, you could resort to this answer which suggests converting your string literal into a concat expression so that you get:

    $name = "Don't d""isplay"
    
    # //q1:Options[q1:Name = concat('Don', "'", 't d"isplay')]
    #                               ^^^^^ use single quotes
    #                                      ^^^ use double quotes
    #                                           ^^^^^^^^^^^^ use single quotes
    

    which you could generate with this:

    $name = "Don't d""isplay"
    
    $squote = "', `"'`", '"
    $expr   = "concat('{0}')" -f $name.Replace("'", $squote)
    
    # Select-Xml -Xml $xml -XPath "//q1:Options[q1:Name = $expr]"
    # ->
    # //q1:Options[q1:Name = concat('Don', "'", 't d"isplay')]
    

    and then the parts of your data that contain double-quotes are delimited with single quotes, and vice versa so they all terminate properly.

    Note - you could probably optimise this for literals without one type of quote or the other and for consecutive single quotes, and it'll need some error handling added for $null and other edge cases, but it basically does the job...


    Update

    Here's a full code sample to show it in action...

    $xml = [xml] "<root><child Name=`"my'name`" /></root>"
    
    $name = "my'name"
    
    $squote = "', `"'`", '"
    $expr   = "concat('{0}')" -f $name.Replace("'", $squote)
    
    Select-Xml -Xml $xml -XPath "//child[@Name = $expr]"
    
    # Node  Path        Pattern
    # ----  ----        -------
    # child InputStream //child[@Name = concat('my', "'", 'name')]