XML parse VBA excel (function trip, & MSXML2.DOMDocument)

Question

I need to parse hundreds of XML files having all the same structure as follows:

<?xml version="1.0" encoding="UTF-8"?>
  <Concepts>
    <ConceptModel name="food">
      <Filters>
        <Filter type="CC"/>
      </Filters>
      <Queries>
        <Query lang="EN">(cheese, bread, wine)</Query>
        <Query lang="DE">(Käse, Brot, Wein)</Query>
        <Query lang="FR">(fromaige, pain, vin)</Query>
      </Queries>
    </ConceptModel>
  </Concepts>

I have read several articles and posts in internet like below but I could not come up with a solution:

• Excel vba Parse Complex XML
• Parse XML File with VBA

So far I am doing:

Dim oXml As MSXML2.DOMDocument
Set oXml = New MSXML2.DOMDocument
oXml.LoadXML ("C:\folder\folder\name.xml")

Dim Queries As IXMLDOMNodeList
Dim Query As IXMLDOMNode

ThisWorkbook.Sheets(3).Cells(i, 1) = "before loop"

Set Queries = oXml.SelectNodes("/concepts/Queries")

MsgBox "how many Queries " &  Queries.Length

For Each Query In Queries
    ThisWorkbook.Sheets(3).Cells(i, 1) = "Works"
    ThisWorkbook.Sheets(3).Cells(i, 2) = Query.SelectNodes("Query").iTem(0).Text
    i = i + 1
Next

This code seems to be understood by VBA but it does not read the contents. The loop does not get read, meaning (I guess) that Queries is not looped at all. Which is confirmed by the fact that the Msgbox "how many queries" gives 0 as result. But actually there are three queries. Could someone give me a hand?

As second issue I would like to ask if

 Dim oXml As MSXML2.DOMDocument

would be the same as

 Dim oXml As MSXML2.DOMDocument60

Since I checked in tools/references "Microsof XML, v6.0"

I thought that the queries having a tag might cause a problem. and I added the follwoing lines:

Dim childs As IXMLDOMNodeList
Set childs = oXml.SelectNodes("/concepts")

MsgBox "childs " & childs.Length

which also gives 0 as result. I would expect 3, since concepts has three children, namely ConceptModel, Filter and Queries. So, I am even more puzzled.

Solution 1

As close as possible to your OP

I 'd draw your attention to several errors or misunderstandings:

• [1] Invalid .LoadXML Syntax

What is then the difference between .LoadXML ("C:\folder\folder\name.xml") and .Load ("C:\folder\folder\name.xml") ?

Load expects a file path and then loads the file content into the oXML object.

LoadXML doesn't expect a file parameter, but its actual XML text content that has to be a well formed string.

• [2] XML distinguishes between lower and upper case, therefore nodes need to be addressed by their exact literal names: the <Query> node wouldn't be identified by "query", "ConceptModel" isn't the same as "conceptmodel".

As second issue I would like to ask if Dim oXml As MSXML2.DOMDocument would be the same as Dim oXml As MSXML2.DOMDocument60, since I checked in tools/references "Microsof XML, v6.0"?

No, it isn't. - Please note that the former declaration would load version 3.0 by default. However it's absolutely preferrable to get the version 6.0 (any other versions are obsolete nowadays!)

As you are using so called early binding (referencing "Microsoft XML, v6.0"), I'll do the same but am referring to the current version 6.0:

Dim oXml As MSXML2.DOMDocument60        ' declare the xml doc object
Set oXml = New MSXML2.DOMDocument60     ' set an instance of it to memory

• [3] misunderstanding some XPath expressions

A starting slash "/" in the XPath expression always refers to the DocumentElement (<Concepts> here), you can add .DocumentElement to your document object instead. A starting double slash "//xyz" would find any "xyz" node if existant.

For instance

    oXml.SelectNodes("//Query").Length 

returns the same childNodes number (here: 3) as

    oXml.DocumentElement.SelectNodes("//Query").Length   ' or 
    oXml.SelectSingleNode("//Queries").ChildNodes.Length ' or even       
    oXml.SelectNodes("/*/*/*/Query").Length`.

Code example with reference to XML version 6.0

Of course you'd have to loop over several xml files, the example only uses one (starting in row 2).

Just for the case of not well formed xml files I added a detailled error Routine that enables you to identify the presumed error location. Load and LoadXML both return a boolean value (True if loaded correctly, False if not).

Sub xmlTest()

Dim ws   As Worksheet: Set ws = ThisWorkbook.Sheets(3)
Dim oXml As MSXML2.DOMDocument60
Set oXml = New MSXML2.DOMDocument60
With oXml
    .validateOnParse = True
    .setProperty "SelectionLanguage", "XPath"   ' necessary in version 3.0, possibly redundant here
    .async = False

    If Not .Load(ThisWorkbook.Path & "\xml\" & "name.xml") Then
        Dim xPE        As Object    ' Set xPE = CreateObject("MSXML2.IXMLDOMParseError")
        Dim strErrText As String
        Set xPE = .parseError
        With xPE
           strErrText = "Load error " & .ErrorCode & " xml file " & vbCrLf & _
           Replace(.URL, "file:///", "") & vbCrLf & vbCrLf & _
          xPE.reason & _
          "Source Text: " & .srcText & vbCrLf & vbCrLf & _
          "Line No.:    " & .Line & vbCrLf & _
          "Line Pos.: " & .linepos & vbCrLf & _
          "File Pos.:  " & .filepos & vbCrLf & vbCrLf
        End With
        MsgBox strErrText, vbExclamation
        Set xPE = Nothing
        Exit Sub
    End If

    ' Debug.Print "|" & oXml.XML & "|"

    Dim Queries  As IXMLDOMNodeList, Query As IXMLDOMNode
    Dim Searched As String
    Dim i&, ii&
    i = 2       ' start row
  ' start XPath  
    Searched = "ConceptModel/Queries/Query"                     ' search string
    Set Queries = oXml.DocumentElement.SelectNodes(Searched)    ' XPath
  ' 
    ws.Cells(i, 1) = IIf(Queries.Length = 0, "No items", Queries.Length & " items")
    ii = 1
    For Each Query In Queries
        ii = ii + 1
        ws.Cells(i, ii) = Query.Text
    Next

End With

End Sub

Additional hints

You also might be interested in an example how to list all child nodes via XMLDOM and to obtain attribute names from XML using VBA.

I include a further hint due to later comment (thanks to @barrowc )

"A further issue with using MSXML, v3.0 is that the default selection language is XSLPatterns instead of XPath. Details on some of the differences between MSXML versions are here and the differences between the two selection languages are discussed here."

In the current MSXML2 version 6.0 XPath 1.0 is fully supported. So it seems XSL Patterns have been implemented by Microsoft in earlier days, basically it can be regarded as a simplified subset of XPath expressions before W3C standardisation of XPath.

MSXML2 Version 3.0 allows the integration of XPath 1.0 at least by explicit selection language setting:

oXML.setProperty "SelectionLanguage", "XPath"   ' oXML being the DOMDocument object as used in original post  

Solution 2

It is the special characters (german alphabet) meaning you need to do something like a batch replace on the XML files so opening line is not this:

<?xml version="1.0" encoding="UTF-8"?>

but this:

<?xml version="1.0" encoding="iso-8859-1" ?>

Code to test with after:

Option Explicit
Public Sub test()
    Dim xmlDoc As Object
    Set xmlDoc = CreateObject("MSXML2.DOMDocument") 'New MSXML2.DOMDocument60
    With xmlDoc
        .validateOnParse = True
        .setProperty "SelectionLanguage", "XPath"
        .async = False
        If Not .Load("C:\Users\User\Desktop\Test.xml") Then
            Err.Raise .parseError.ErrorCode, , .parseError.reason
        End If
    End With
    Debug.Print xmlDoc.SelectNodes("//Query").Length
End Sub

---

This is the XML I am using:

<?xml version="1.0" encoding="iso-8859-1" ?>
  <Concepts>
      <ConceptModel name="food">
    <Filters>
      <Filter type="CC"/>
    </Filters>
    <Queries>
      <Query lang="EN">(cheese, bread, wine)</Query>
      <Query lang="DE">(Käse, Brot, Wein)</Query>
      <Query lang="FR">(fromaige, pain, vin)</Query>
   </Queries>
  </ConceptModel>
</Concepts>

Comments

[JFerro]

but that would mean that if there is no "ü", "ä" or whatever of this stily it should work. And it doesnt. Unfortunately I can not influence the content of the files themselves, I can only read data.

[QHarr]

I'm not sure but I can assure you that it does work with the above change. There maybe other ways to achieve the same such as reading in using a different method first and parsing that to the xmldocument.

[JFerro]

Just for me to understand. I did the following. I opened the xml file in a text editor and changed (as you said) the head into: <?xml version="1.0" encoding="iso-8859-1" ?> save it and run my code again. same result. Is that what you mean?

[JFerro]

Another try: I save an XML file from a text editor without any head specifing encoding and save it with XML extension as text file. (and without german special characteres). Did not work neither

[JFerro]

Now I did. Yours work. Not only that, I change the initial like by the one containing UFT and then gave error.

[JFerro]

So thanks. I think I have to try other methods then, like reading pure text and parsing. I can not change the files since they are read by other programs as well.

[JFerro]

What is then the difference between .LoadXML ("C:\folder\folder\name.xml") and .Load ("C:\folder\folder\name.xml") ??? I pasted my file in "XMLvalidator" and there was no errors, so, what do you mean exactly by: "<< CANNOT load this - NO valid XML string input" thx

[T.M.]

Give me some time, I'll edit this and your versioning question in my answer. Hadn't the time to study your post completely as you already accepted @QHarris ' valid approach.

[JFerro]

of course, all the time of the wrold,... Harris approach is "a" solution, but of course only if I could change the files themselves. I can not touch the files. I pasted my file in "XMLvalidator" and there was no errors, so, what do you mean exactly by: "<< CANNOT load this - NO valid XML string input" Carefully looking at your code it looks like changing in my code .loadXML by .load("... should work, but it doesnt. Since pasting your condition if not .load("... does not cause Fehler, but still the rest does not work

[T.M.]

BTW a) Do you want to get all <Query> items or just the English one, i.e. .Item(0)? In this case you needn't loop through the whole nodelist, you can pick the wanted item via index if you get a nodelist length greater than 0. b) Are there other <Queries> subnodes or does each XML file consist only of one such node and e.g. three language subnodes?

[JFerro]

all of them. ALL the XML are exactly the same, only changes the name of the file, and hence also <ConceptModel name="XXXXXX">, and the content of every of the Query. Which are always EN, DE, FR

[T.M.]

Tried to answer your questions. Of course you'll have to change to .Load(filepath) if you refer to a file path otherwise you don't get xml content - (cf. `Debug.Print oXML.xml). Furthermore your XPath settings searched for nodes that couldn't be identified, therefore you got no results :-)

[barrowc]

A further issue with using MSXML, v3.0 is that the default selection language is XSLPatterns instead of XPath. Details on some of the differences between MSXML versions are here and the differences between the two selection languages are discussed here

[T.M.]

Thank you for the additional links as well as stating the differences more precisely; I only commented in code that an explicit .setProperty "SelectionLanguage", "XPath" is necessary in Version 3 - may I include that in my answer? - @barrowc

[barrowc]

You are more than welcome to include any part of my comment in the answer