Confused about getting data from an XML file 1

[PPettit]

I've got a script that checks XML files for errors. One thing that it does is to make sure that there is an attachment file name present. I finally figured out how to make it perform as expected, but I was hoping that someone could help me to understand what I was doing wrong.

In essence, the pertinent XML structure is like so:

<Invoice>
<filename_attachment pagecount="1">408901.pdf</filename_attachment>
</Invoice>

Occasionally, I need to process old files that are slightly different (no pagecount attribute):

<Invoice>
<filename_attachment>370000.pdf</filename_attachment>
</Invoice>

Using this:
$xmlFilename_Attachment = [string]$xd.DocumentElement.filename_attachment.InnerText
Old XML returned nothing
New XML returned "408901.pdf"


If I cut off ".InnerText"
Old XML returned "370000.pdf"
New XML returned "System.Xml.XmlElement"

I finally settled on using this:

$xmlFilename_Attachment = $xd.get_DocumentElement().GetElementsByTagName("filename_attachment").Item(0).InnerText

It seems to work on either version of my XML files.

I'd like to have a better understanding as to why the inclusion/exclusion of ".InnerText" and the "pagecount" attribute screwed things up. Also, I'd like to know if there is a better way to get the file name information.

 

[tsuji]

That can be something that makes users "mad" about. ps seems to have its own simplified native object model (based on dot-notation) on top of what .net framework is prepared to support.

The clue to determine what happens is to see the type. Suppose $xd holds the document with attribute pagecount whereas $xd_old holds without, the "old" kind of document.

[1] $xd
>[tt]$xd.Invoice.filename_attachment.gettype().fullname
[blue]Sytem.Xml.XmlElement[/blue]
[/tt]
And the two data can be retrieved like this (amongst some others).
[tt]
>$xd.Invoice.filename_attachment."#text"
[blue]408901.pdf[/blue]
>$xd.Invoice.filename_attachment.pagecount
1
[/tt]
[2] $xd_old
[tt]>$xd_old.Invoice.filename_attachment.gettype().fullname
[blue]System.String[/blue]
[/tt]
Since it is System.String type, InnerText is not one of its member property! Its "innertext" is in fact directly from that reference.
[tt]>$xd_old.Invoice.filename_attachment
[blue]370000.pdf[/blue]
[/tt]
[3] That is distressing as it seems to build for the purpose of making users easier to process the document and it finishes with making the matter quite a bit of a mess. It seems users should stict to what .net framework support to make the reasoning more consistent.

[4] Furthermore, the model exposed using the dot-notation seems to take the tag names case-insensitive, Invoice, invoice and filename_attachment, Filename_Attachment etc are all accepted graciously. I am not sure it is wise, though.

 

[PPettit]

Thanks, tsuji. I'm going to try and stick to what .Net supports. That seems like the best approach unless I'm just doing something quick from the command line.