XML encoding is not added

I'm not really used to using MSXML (for Dephi I use a third party XML library), but it seems the encoding attribute is only added if you set it to something different then UTF-8. If I change your code to:

  Doc := NewXMLDocument;
  Doc.Version  := '1.0';
  Doc.Encoding := 'utf-16';

Select allOpen in new window

the encoding attribute is added, it almost looks like default MSXML behaviour. Why do you need it to be there?

(en de groeten :p)

Please note - activate xml document first:

var
  Xml: IXmlDocument;
begin
  CoInitialize(nil);
  try
    Xml := NewXMLDocument;
    Xml.Active := True;
    Xml.Options := Xml.Options + [doNodeAutoIndent];
    Xml.ParseOptions := Xml.ParseOptions + [poPreserveWhiteSpace];
    Xml.Encoding := 'UTF-8';
    ShowMessage(Xml.XML.Text);
  finally
    CoUninitialize;
  end;

Select allOpen in new window

@sinisav, check the implementation of NewXMLDocument()

function NewXMLDocument(Version: DOMString = '1.0'): IXMLDocument;
begin
  Result := TXMLDocument.Create(nil);
  Result.Active := True;
  if Version <> '' then
    Result.Version := Version;
end;

Select allOpen in new window

As you can see, it sets active to true for you :)

When I test your code here, there is still no encoding attribute added.

MerlijnB: The XML must be build up in a certain way, as it is described in the SEPA documentation. They apparently require the encoding.

Sinisav: Your code gives the same result here:

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

It's a strange behaviour. Even when I manually add the encoding it still drops of after I load the string to XML.

    sDoc := Doc.XML.Text;
    sDoc := StringReplace(sDoc, 'encoding="utf-16"', 'encoding="utf-8"', [rfReplaceAll]);
    Doc := LoadXMLData(sDoc);

Select allOpen in new window

Found this code in XMLDoc:

function XMLStringToUnicodeString(const XMLString: AnsiString): UnicodeString;
const
  AnsiEncodings: array[0..1] of string = ('ISO-8859-1', 'US-ASCII');
  UnicodeEncodings: array[0..2] of string = ('UTF-16', 'UCS-2', 'UNICODE');
var
  Encoding: string;
begin
  Encoding := ExtractAttrValue(SEncoding, Copy(string(XMLString), 1, 50), '');
  { No Encoding is assumed to be UTF-8 }
  if (Encoding = '') or SameText(Encoding, 'UTF-8') then
    Result := UTF8ToString(XMLString)
  { Latin1 and ASCII can use standard AnsiString->UnicodeString conversion }
  else if EncodingMatches(Encoding, AnsiEncodings) then
    Result := string(XMLString)
  else
  { All others fail }
    XMLDocError(SUnsupportedEncoding, [Encoding]);
  CheckEncoding(Result, UnicodeEncodings);
end;

Select allOpen in new window

So I guess it is the default XML-encoding as MerijnB has mentioned.
It is still strange that I was able to add the 'utf-8' encoding before???

@DelphiWizard

for some reason this seems to be a workaround:

var xml: IXMLDocument;
    s: utf8string;
begin
 Xml := NewXMLDocument;
 Xml.Options := Xml.Options + [doNodeAutoIndent];
 Xml.ParseOptions := Xml.ParseOptions + [poPreserveWhiteSpace];
 Xml.Encoding := 'UTF-8';
 XML.SaveToXML(s);
end;

Select allOpen in new window

Note that s is an utf8string! If you change it to a normal string it stops working.

Alternatively you could check out 3th party libraries (like SimDesign's TNativeXML: http://www.simdesign.nl/xml.html)

MerijnB: Which external XML library do you use?

I guess it depends on delphi version. In newer delphi strings are unicode (and not utf8), so there is no encoding in your app, but I have it (D6). So, this utf8string is good as is ansistring too in your example.

@DelphiWizard, we use TNativeXML

MerijnB: Your code produces the same result.
I guess    Doc := LoadXMLData(sDoc);
Removes the utf-8 encoding in the end.

Why do you call LoadXMLData() ?

To get the XML, just do  XML.SaveToXML(s), where s is an UTF8string

After building the whole XMl, I need to replace a string.

var sDoc : UTF8String;
begin
    sDoc := Doc.XML.Text;
    sDoc := StringReplace(sDoc, '<CstmrDrctDbtInitn xmlns="">', '<CstmrDrctDbtInitn>', [rfReplaceAll]);
    Doc := LoadXMLData(sDoc);
    Doc.SaveToFile(ExportMap + 'SEPA XML ' + FormatDateTime('yyyy-mm-dd',Date) + '.xml');
end;

Select allOpen in new window

Where do you assign content to var "s"?

Looks good, however this is not working.

    // get the XML data in a utf-8 string to force utf-8 attribute
    sTemp := Doc.SaveToXML(sTemp);

Incompatible types: 'UTF8String' and 'procedure, untyped pointer or untyped parameter'

Perfect, thank you very much.

To be complete:
Creating the file can be as easy as:

TFile.WriteAllText('c:\somefile.xml', sDoc);  // uses IOUtils

Select allOpen in new window