Trouble with (very) simple XML-parsing

[Tommi Prami 130]

(I use 10.3.1)

 

If got an code :

 

uses
  Winapi.msxml, Winapi.MSXMLIntf;

procedure TForm10.Button1Click(Sender: TObject);
var
  LDocument: IXMLDomDocument3;
  LCurrrentValueNode: IXMLDOMNode;
  LNodeList: IXMLDOMNodeList;
  I: Integer;
begin
  LDocument := CoDOMDocument60.Create;
  LDocument.async := False;
  LDocument.validateOnParse := False;
  LDocument.ResolveExternals := False;
  LDocument.PreserveWhiteSpace := True;

  LDocument.loadXML(Memo1.Lines.Text);
  if LDocument.parseError.errorCode <> 0 then
    raise Exception.CreateFmt('XML parsing error : %s at row %d position %d',
      [LDocument.parseError.reason, LDocument.parseError.Line, LDocument.parseError.linepos]);

  if Assigned(LDocument.documentElement) then
    LNodeList := LDocument.documentElement.getElementsByTagName('IBAN')
  else
    LNodeList := LDocument.getElementsByTagName('IBAN');

  for I := 0 to LNodeList.Length - 1 do
  begin
    LCurrrentValueNode := LNodeList.Item[I];
    Memo2.Lines.Add(LCurrrentValueNode.nodeValue);
  end;

  repeat
    LCurrrentValueNode := LNodeList.nextNode;
    if Assigned(LCurrrentValueNode) then
      Memo2.Lines.Add(LCurrrentValueNode.nodeValue)
    else
      Break;
  until True;
end;

and XML

 

<?xml version="1.0" encoding="UTF-8"?>
<Document xmlns="urn:iso:std:iso:20022:tech:xsd:camt.053.001.02">
  <BkToCstmrStmt>
    <Stmt>
      <Acct>
        <Id>
          <IBAN>FI0434270410003403</IBAN>
        </Id>
      </Acct>
    </Stmt>
    <Stmt>
      <Acct>
        <Id>
          <IBAN>FI7542316072000169</IBAN>
        </Id>
      </Acct>
    </Stmt>
  </BkToCstmrStmt>
</Document>

Any idea why getElementsByTagName does not return Tags it should (I think).

 

Once you're set to do something pretty simple and nothing works. I remember now why I do not enjoy the M$Xml too much.

 

Also if you can recommend better replacement XML libraries, feel free to do so.

Edited March 15, 2019 by Tommi Prami

[Tommi Prami 130]

With OmniXML dom I got slightly closer

 

procedure TForm10.Button2Click(Sender: TObject);
var
  LXMlDocument: IXMLDocument;
  LDomList: IDOMNodeList;
  LValueNode: IDOMNode;
  I: Integer;
begin
  DefaultDOMVendor := sOmniXmlVendor;
  LXMlDocument := LoadXMLData(Memo1.Lines.Text);
  LDomList := LXMlDocument.DOMDocument.documentElement.getElementsByTagName(TAG_IBAN);

  for I := 0 to LDomList.length - 1 do
  begin
    LValueNode := LDomList[I];
    Memo2.Lines.Add(LValueNode.nodeName + ': ' + LValueNode.nodeValue);
  end;
end;

it find the tags, but nodeValue returns empty. 

 

-Tee-

[Der schöne Günther 316]

Doesn't getElementsByTagName(..) only list direct descendants? You will either have to search recursively, or use something like an XPath expression to get all the nodes you are interested in.

[Tommi Prami 130]

At least all the documentation I've found tells that it should. find others also. m$xml docs and also more generic DOM ones.

 

What ever is the implementation is totally different story.

 

-Tee-

[Anders Melander 1782]

1 hour ago, Tommi Prami said:

if Assigned(LDocument.documentElement) then 
  LNodeList := LDocument.documentElement.getElementsByTagName('IBAN')
 else
   LNodeList := LDocument.getElementsByTagName('IBAN');

 

Why do you have this test? I would think that the second variant (LDocument.getElementsByTagName) would be good enough...

 

I also think it would be safer if you used XPath or specified the path in your search criteria:

LDocument.getElementsByTagName('/BkToCstmrStmt/Stmt/Acct/Id/IBAN');

 

[Tommi Prami 130]

First test was there just to get an idea is there one and which it'll use. IT seems to depend on something. Done so many tests so can't really give specifics on that.

 

At least that xpath down work, because structure of XML is not always same. It has to be more structure agnostic.

 

 

[Tommi Prami 130]

Did test with Xml.XMLIntf.pas IXMLDocument also. OmniXML as vendor or not.

 

Made own recursive routine, based on OmniXML implementation, it'll find nodes just fine, but not node Values, which seems to be super weird.

 

function GetElementsByTagName(const ADocumentElement: IDOMElement; const AElementName: string): TList<IDOMNode>;

  procedure InternalGetElementsByTagName(const ANode: IDOMNode; const AElementName: string;
    const ANodeList: TList<IDOMNode>);
  var
    I: Integer;
    LChildNode: IDOMNode;
  begin
    if ANode.HasChildNodes then
      for I := 0 to ANode.ChildNodes.Length - 1 do
      begin
        LChildNode := ANode.ChildNodes.Item[I];
        if (LChildNode.NodeType = ELEMENT_NODE) and ((LChildNode as IDOMElement).NodeName = AElementName) then
          ANodeList.Add(LChildNode);
        InternalGetElementsByTagName(LChildNode, AElementName, ANodeList);
      end;
  end;

begin
  Result := TList<IDOMNode>.Create;

  InternalGetElementsByTagName(ADocumentElement, AElementName, Result);
end;

 

Edited March 15, 2019 by Tommi Prami