XML parsing, TXMLDocument

Question

I have a problem with parsing XML.

How to get field values se_url and phrase?
I need to get link1_1, link1_2, key1, link2_1, link2_2, key2... which are in se_url and phrase.

I did not find in Google how to do it (also did not find a manual on how to work with TXMLDocument).

<doc>
  <date2>20120214</date2>
  <date1>20120214</date1>
  <data count="116">
    <row>
      <search_engines count="2">
        <search_engine>
          <se_url>link1_1</se_url>
          <se_page>1</se_page>
          <se_id>2</se_id>
        </search_engine>
        <search_engine>
          <se_url>link1_2</se_url>
          <se_page>1</se_page>
          <se_id>3</se_id>
        </search_engine>
      </search_engines>
      <denial>0.4889</denial>
      <visits>45</visits>
      <page_views>52</page_views>
      <phrase>key1</phrase>
      <visit_time>126</visit_time>
      <depth>1.1556</depth>
    </row>
    <row>
      <search_engines count="2">
        <search_engine>
          <se_url>link2_1</se_url>
          <se_page>1</se_page>
          <se_id>3</se_id>
        </search_engine>
        <search_engine>
          <se_url>link2_2</se_url>
          <se_page>1</se_page>
          <se_id>6</se_id>
        </search_engine>
      </search_engines>
      <denial>0.5714</denial>
      <visits>42</visits>
      <page_views>50</page_views>
      <phrase>key2</phrase>
      <visit_time>109</visit_time>
      <depth>1.1905</depth>
    </row>
  </data>
</doc>

Answer 1

Try this:

uses ComObj, MSXML;

procedure TForm1.Button1Click(Sender: TObject);
var
  xml: IXMLDOMDocument;
  node: IXMLDomNode;
  nodes_row, nodes_se: IXMLDomNodeList;
  i, j: Integer;
  url: string;
begin
  // put url or file name
  url := 'http://softez.pp.ua/gg.xml';

  xml := CreateOleObject('Microsoft.XMLDOM') as IXMLDOMDocument;
  xml.async := False;
  xml.load(url); // or use loadXML to load XML document using a supplied string
  if xml.parseError.errorCode <> 0 then
    raise Exception.Create('XML Load error:' + xml.parseError.reason);

  Memo1.Clear;
  nodes_row := xml.selectNodes('/doc/data/row');
  for i := 0 to nodes_row.length - 1 do
  begin
    node := nodes_row.item[i];
    Memo1.Lines.Add('phrase=' + node.selectSingleNode('phrase').text);
    nodes_se := node.selectNodes('search_engines/search_engine/se_url');
    for j := 0 to nodes_se.length - 1 do
    begin
      node := nodes_se.item[j];
      Memo1.Lines.Add('url=' + node.text);
    end;
    Memo1.Lines.Add('--------------');
  end;
end;

Result:

phrase=key1
url=link1_1
url=link1_2
--------------
phrase=key2
url=link2_1
url=link2_2
--------------

A Reference to IXMLDOMDocument

Answer 2

If you first include these 3 general purpose library routines ....

uses XMLDoc, XMLIntf, xmldom;

function CreateXMLDocument( var Owner1: TComponent): TXMLDocument;
begin
Owner1 := TComponent.Create( nil);
result  := TXMLDocument.Create( Owner1);
result.Options := [doNodeAutoCreate, doNodeAutoIndent, doAttrNull,
                   doAutoPrefix, doNamespaceDecl];
result.DOMVendor := GetDOMVendor( 'MSXML');
end;

function XPATHSelect( const FocusNode: IXMLNode; const sXPath: string): TArray<IXMLNode>;
var
  DomNodeSelect: IDomNodeSelect;
  DOMNode      : IDomNode;
  DocAccess    : IXmlDocumentAccess;
  Doc          : TXmlDocument;
  DOMNodes     : IDOMNodeList;
  iDOMNode     : integer;
begin
SetLength( result, 0);
if assigned( FocusNode) and
   Supports( FocusNode.DOMNode, IDomNodeSelect, DomNodeSelect) then
    DOMNodes := DomNodeSelect.SelectNodes( sXPath);
if not assigned( DOMNodes) then exit;
SetLength( result, DOMNodes.Length);
for iDOMNode := 0 to DOMNodes.Length - 1 do
  begin
  Doc := nil;
  DOMNode := DOMNodes.item[iDOMNode];
  if Supports( DOMNode, IXmlDocumentAccess, DocAccess) then
    Doc := DocAccess.DocumentObject;
  result[ iDOMNode] := TXmlNode.Create( DOMNode, nil, Doc) as IXMLNode;
  end
end;


function XPATHSelectFirst( const FocusNode: IXMLNode; const sXPath: string; var SelectedNode: IXMLNode): boolean;
var
  DomNodeSelect: IDomNodeSelect;
  DOMNode      : IDomNode;
  DocAccess    : IXmlDocumentAccess;
  Doc          : TXmlDocument;
begin
SelectedNode := nil;
if assigned( FocusNode) and
   Supports( FocusNode.DOMNode, IDomNodeSelect, DomNodeSelect) then
  DOMNode := DomNodeSelect.selectNode( sXPath);
if assigned( DOMNode) and
   Supports( DOMNode.OwnerDocument, IXmlDocumentAccess, DocAccess) then
  Doc := DocAccess.DocumentObject;
if Assigned( DOMNode) then
  SelectedNode := TXmlNode.Create( DOMNode, nil, Doc);
result := assigned( SelectedNode)
end;

Then A much neater solution is ...

procedure TForm2.btn1Click(Sender: TObject);
const
  DocumentSource =  'http://softez.pp.ua/gg.xml';
var
  Doc: IXMLDocument;
  DocOwner: TComponent;
  RowNode, PhraseNode, UrlNode: IXMLNode;

  procedure PutLn( const LineFmt: string; const Args: array of const);
  begin
  memo2.Lines.Add( Format( LineFmt, Args))
  end;

begin
memo2.Clear;
Doc := CreateXMLDocument( DocOwner);
Doc.LoadFromFile( DocumentSource);
for RowNode in XPATHSelect( Doc.DocumentElement, '//row[phrase]') do
  begin
  if not XPATHSelectFirst( RowNode, 'phrase', PhraseNode) then continue;
  PutLn( 'phrase=%s', [PhraseNode.NodeValue]);
  for UrlNode in XPATHSelect( RowNode, 'search_engines/search_engine/se_url') do
    PutLn( 'url=%s', [UrlNode.NodeValue]);
  PutLn('--------------',[])
  end;
DocOwner.Free;
end;

This was tested on Delphi 2010 and works a treat.