OmniXML support forum

[Help] Re: Traversing an XML file

gabr — Mon, 28 Jul 2008 18:53:30 +0200

Text (node data) is represented in a XML DOM as a node with name #text and type TEXT_NODE (integer 3). OmniXML.pas defines following node types: const // element node ELEMENT_NODE = 1; // attribute node ATTRIBUTE_NODE = 2; // text node TEXT_NODE = 3; // CDATA section node CDATA_SECTION_NODE = 4; // entity reference node ENTITY_REFERENCE_NODE = 5; // entity node ENTITY_NODE = 6; // processing instruction node PROCESSING_INSTRUCTION_NODE = 7; // comment node COMMENT_NODE = 8; // document node DOCUMENT_NODE = 9; // document type node DOCUMENT_TYPE_NODE = 10; // document fragment node DOCUMENT_FRAGMENT_NODE = 11; // notation node NOTATION_NODE = 12; Most probably you're only interested in types ELEMENT_NODE and ATTRIBUTE_NODE. As an example, following code from OmniXMLUtils finds first child node of type TEXT_NODE to extract node text: function GetTextChild(node: IXMLNode): IXMLNode; var iText: integer; begin Result := nil; for iText := 0 to node.ChildNodes.Length-1 do if node.ChildNodes.Item[iText].NodeType = TEXT_NODE then begin Result := node.ChildNodes.Item[iText]; break; //for end; end; { GetTextChild }

[Help] Traversing an XML file

Stefano Bragaglia — Mon, 28 Jul 2008 18:30:13 +0200

I'd like to read an XML file and take different actions depending on the name of the nodes I read... I don't know the structure of the file in advance, so I have to "check out" the structure node by node... I read the root node of the document with: Root := XMLDocument.DocumentElement; I'd like to iterate through all the children of the root and, recursively, to all the children of the children and so on... I tryed something like: Current := Root.FirstChild; while Assigned(Current) do begin // Take an action depending on name and recurse Current := Current.NextChild; end; or similarly with Root.ChildNodes.Reset; Current := Root.ChildNodes.NextNode; while Assigned(Current) do begin // Take an action depending on name and recurse Current := Current.NextChild; end; but it doesn't work as I expect... the child's node name I read is always "#text" (without quotes) no matter the real name is... What am i doing wrong? Can someone point me out the right code I should use? Thanks in advance, best wishes, Stefano Bragaglia

[Bug Reports] LoadFromStream doesn't work

Lee — Tue, 22 Jul 2008 05:05:04 +0200

Hi, I've tried several times to load an xml document from a stream and it doesn't work. Example: ADoc := CreateXMLDoc; AStream.Seek(0, soFromBeginning); ADoc.LoadFromStream(AStream); //

[Help] Re: Self-closing tags

mremec — Thu, 26 Jun 2008 16:04:09 +0200

I am not sure I understand your question. However, XML tags are always closed automatically. Parser must understand both forms but it uses short form as output.

[Help] Self-closing tags

Mark — Mon, 23 Jun 2008 15:28:28 +0200

How can I set tags to be self-closing: ie: becomes or to Cheers

[Bug Reports] Re: More support needed for xPath

gabr — Thu, 13 Mar 2008 08:06:17 +0100

Noted, but currently we have no plans for extending xPath.

[Bug Reports] More support needed for xPath

delphiForever — Thu, 13 Mar 2008 00:07:28 +0100

Hi, if you do //[@withthistag] You should get all the nodes 'withthistag' if you do if you do //[@withthistag and @andthisone] you should get all nodies 'withthistag' 'andthisone' Yet, that isn't working. also or isn't working and //[@withthistag="monkey" and @andthisone="ape"] should work with and/or/not/ ... I used to use XML partners and looking for a replacement becouse xmlparners doesn't work under delphi 2007, wich sux. :| thanx. Kind regards.

[Announcements] OmniXML release 2008-02-15

mremec — Fri, 15 Feb 2008 10:03:46 +0100

I am pleased to announce that new release of OmniXML library is available. Big thanks to all contributors and users for feedback, bugfixes and general support. What's new in this release: * OmniXML.pas, OmniXML_Dictionary.pas: Added dictionary to store tag names, attribute names and attribute values. * OmniXML.pas: Fixed node type in IXMLDocumentType (reported by qrbn). * OmniXML.pas: Fixed bug in FindCharReference (thanks to Andrej Zavrsnik for bug report and fix). * OmniXMLXPath.pas: Fix: Parser was calling 8-bit versions of Pos and PosEx. Fixed PosEx to accept 16-bit parameters and added 16-bit version of Pos. (gp) * OmniXML_Dictionary.pas: Delphi 7 compatibility fix. * OmniXMLXPath.pas: Deleted reference to (not used) StrUtils unit. For a list of all changes please see ChangeLog.txt. To compile OmniXML with Delphi 2005 or newer, use OmniXMLCore.dpk. This single package compiles with Delphi 2005 or newer. It defines LIBSUFFIX compiler directive depending on compiler version. As a result, after compile with Delphi 2006, you'll get OmniXMLCore.dcp and OmniXMLCore_BDS4.bpl files.

[Help] Re: Problems installing

mremec — Fri, 15 Feb 2008 09:18:05 +0100

OmniXMLXPath.pas references StrUtils which appeared in Delphi 6. However, it seems that the reference is not needed because test program compiles without any errors even if I delete this reference. So, open OmniXMLXPath.pas, search for StrUtils and delete it. Then, it should compile in D5 also.

[Help] Re: Problems installing

VM — Fri, 15 Feb 2008 02:36:47 +0100

upon installing the package at DELPHI 5, it keep on lookin the strutils.dcu Where i can find it. thanks venn

[Help] Re: How access directly to these nodes ?

Bruno — Thu, 24 Jan 2008 23:52:57 +0100

oh yess ! Thank's a lot Sincerely, Bruno

[Help] Re: How access directly to these nodes ?

gabr — Thu, 24 Jan 2008 15:25:30 +0100

Better: var node: IXMLNode; node := SelectNode(nodeList.Item[iNode], './name/fullName'); if assigned(node) then fullName := GetTextChild(node).Text else fullName := '';

[Help] Re: How access directly to these nodes ?

Bruno — Thu, 24 Jan 2008 15:12:51 +0100

Just for information, the GetTextChild do an error if the node doesn't exists. So I use : Try fullName := GetTextChild(SelectNode(nodeList.Item[iNode], './name/fullName')).Text; Except fullname:=''; End;

[Help] Re: How access directly to these nodes ?

Bruno — Wed, 23 Jan 2008 18:34:26 +0100

You are GREAT ! Many thank's for your quick reply ! Bruno

[Help] Re: How access directly to these nodes ?

gabr — Wed, 23 Jan 2008 08:02:16 +0100

The simplest way is to use OmniXMLUtils and do for iNode := 0 to nodeList.Length - 1 do begin fullName := GetTextChild(SelectNode(nodeList.Item[iNode], './name/fullName')).Text; country := GetTextChild(SelectNode(nodeList.Item[iNode], './address/country')).Text; end;

[Help] How access directly to these nodes ?

Bruno — Tue, 22 Jan 2008 23:58:15 +0100

Hi all, I've this XML structure, and I use D7. I try to extract directly the fullname and the Country. Is it a way to do that directly ? ..... Ericsson Inc. Research Triangle Park NC US UNITED STATES I've done : .... nodeList := XPathSelect(CurrentNode, './bibliographicInformation/assignees/assignee'); if nodeList.Length>0 then for iNode:=0 to nodeList.Length-1 do Begin // WHAT MUST I DO HERE TO EXTRACT FULLNAME AND COUNTRY ? end; ...... Many thank's for your help Bruno

[Bug Reports] Re: The unicode surrogate code points are not handled

alphax — Sat, 12 Jan 2008 13:44:06 +0100

I am sorry, the last statement should be changed to: if not Doc.Load(FN) then raise Exception.Create('Parse document failed');

[Help] Re: Getting the value

Richard S — Sat, 12 Jan 2008 01:03:12 +0100

Thank you very much indeed, that does exactly what I want.

[Help] Re: Getting the value

gabr — Fri, 11 Jan 2008 14:38:42 +0100

The simplest way: uses OmniXML, OmniXMLUtils; var xml: IXMLDocument; ... load the xml s := GetTextChild(xml.SelectSingleNode('/Unit')).Text; // 'Part Ground Floor' s := GetTextChild(xml.SelectSingleNode('/Unit/Area')).Text; // 'Retail'

[Help] Getting the value

Richard S — Fri, 11 Jan 2008 13:56:25 +0100

Hi - probably a silly basic question, but what's the best way to extract the values "Part Ground Floor" from and "Retail" from from the snippet below? I find getting attributes very straightforward, but extracting the values is currently defeating me. {BFB68548-81CA-4E90-A34A-B7D2324FE5CE} Part Ground Floor {631A2448-7315-4183-924D-B9CFBD410A9F} Retail

[Bug Reports] Re: OmniXMLPath.XPathSelect() implementation should NOT use Ansi edtion Pos()

gabr — Thu, 10 Jan 2008 18:02:19 +0100

I fully agree - this is a bug. I'll fix the CVS version today.

[Bug Reports] OmniXMLPath.XPathSelect() implementation should NOT use Ansi edtion Pos()

alphax — Thu, 10 Jan 2008 13:51:05 +0100

The methods TXMLXPathEvaluator.GetNextExpressionPart() and TXMLXPathEvaluator.SplitExpression() both are use Ansi-edition Pos() function for locate the char index, but the input parameter are WideStrings, so, the postion returned by Pos() could not be used to access the the wide string, if some multibytes character contained in that wide string. for example: endElement := PosEx('/', FExpression, FPosExpression); ... element := Copy(FExpression, FPosExpression, endElement - FPosExpression); endElement just the position of the ansi-edition of the FExpression, if FExpression contains multibytes character(in the attribute value expression), the Copy() call in the next statement will return incorrect result. So, I think, all Pos() call and all PosEx() call should be replaced to the wide-edition call. TXMLXPathEvaluator.PosEx() should be declared as: function PosEx(ch: WideChar; const s: WideString; offset: integer = 1): integer; and we should introduce a function TXMLXPathEvaluator.Pos(Ch: WideChar; const S: WideString): Integer;

[Bug Reports] Re: The unicode surrogate code points are not handled

alphax — Wed, 9 Jan 2008 06:35:15 +0100

Ok, I am make a sample code which can reproduce this problem. //Code copy from OpenXML\Utilities\cUnicodeCodecs.pas and modified procedure UCS4CharToUTF8(const Ch: UCS4Char; const Dest: Pointer; const DestSize: Integer; out SeqSize: Integer); var P : PByte; begin P := Dest; if Ch < $80 then // US-ASCII (1-byte sequence) begin SeqSize := 1; if not Assigned(P) or (DestSize

[Bug Reports] Re: The unicode surrogate code points are not handled

mremec — Wed, 26 Dec 2007 10:54:02 +0100

I'm sorry, I somehow forgot to answer your first mail. Can you please send a sample document that shows this problem to mremec@gmail.com? Many thanks in advance.

[Bug Reports] Re: The unicode surrogate code points are not handled

alphax — Tue, 25 Dec 2007 13:24:36 +0100

Has any plan for fix this problem?

[Help] Re: GetElementsByTagName

Peter Gore — Thu, 20 Dec 2007 17:06:04 +0100

Thanks Miha, that's a lot clearer. Regards Peter Gore

[Help] Re: GetElementsByTagName

mremec — Thu, 20 Dec 2007 08:41:23 +0100

NodeName is a property of IXMLNode and all child interfaces. IXMLElement defines additional property TagName, that returns the same value as NodeName property. Function GetElementsByTagName does what its name tells: it returns elements (IXMLElement) by tag names. Please open http://www.w3.org/TR/2000/WD-DOM-Level-1-20000929/level-one-core.html#ID-1950641247 and navigate one or two pages down to table that describes nodeName, nodeValue and attributes return values according to node type. There you will see that NodeName for Element (IXMLElement) returns tag name. And yes, line 2657 could (should) be better as follows: if (ChildNode.NodeType = ELEMENT_NODE) and ((ChildNode as IXMLElement).TagName = Name) then Result.AddNode(ChildNode);

[Help] GetElementsByTagName

Peter Gore — Thu, 20 Dec 2007 08:07:26 +0100

This function actually seems to evaluate against NodeName rather than TagName. Line 2657 of OmniXML refers if (ChildNode.NodeType = ELEMENT_NODE) and ((ChildNode as IXMLElement).NodeName = Name) then Result.AddNode(ChildNode); TagName in my testing is read only and always equal to the root Element Name. Should it therefore not be GetElementsByNodeName? Regards Peter Gore

[Help] Re: IXMLNode and IXMLElement

Peter Gore — Thu, 20 Dec 2007 05:48:03 +0100

OK thanks a lot

[Help] Re: IXMLNode and IXMLElement

mremec — Wed, 19 Dec 2007 20:02:06 +0100

In short, IXMLElement inherits from IXMLNode and defines additional properties and methods.