We Keep Coding

How do I get a serialized object to look the same as it's original source xml file?

I have a xsd file, defined by an external company, that I used with xsd.exe to generate classes. I can use a provided xml file to deserialize into an object using the generated classes just fine, but there are a few cases where I need to have smaller portions of the xml as a XDocument. I won't know the path in these portions until run time, so I'm using the xml for:
XElement element = xml.XPathSelectElement(path);
The issue I'm having is that serialized result doesn't match the incoming xml quite right, which makes the select return null. How do I get a serialized object to look like the incoming file? Did I possibly generate the classes incorrectly with xsd.exe? I'll eventually need to use the same generated code to generate my own xml files.
Here's the code I'm currently using to serialize
var xml = new XDocument();
using (var writer = xml.CreateWriter())
{
List<Type> known = new List<Type>();
known.Add(typeof(ObjType1));
...
var serializer = new DataContractSerializer(typeof(Detail), known);
serializer.WriteObject(writer, sourceDetailObj);
}
The serialized result:
<Detail xmlns:i="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://schemas.datacontract.org/2004/07/CustomNameSpace">
...
<numberField>1</numberField>
<detailTypeField>
<objField i:type="ObjType1">
<valObjField i:nil="true" />
...
</objField>
</detailTypeField>
...
</Detail>
What it should look like:
<Detail>
...
<Number>1</Number>
<DetailType>
<ObjType1>
...
</ObjType1>
</DetailType>
...
</Detail>
Here's one of the classes xsd generates:
public partial class DetailType {
private object objField;
[System.Xml.Serialization.XmlElementAttribute("ObjType1", typeof(ObjType1))]
...
public object Obj {
get {
return this.objField;
}
set {
this.objField = value;
}
}
}
Obj can be one of several classes.

The problem with using a DataContractSerializer is that it is optimized for sending messages between WCF services and won't necessarily produce the same "classic" xml that the XmlSerializer does.
In particular, XmlSerializer will serialize all public members unless you tell it not to, but for DataContractSerializer it won't serialize unless you tell it to. This was done to help make WCF faster; you only get what you ask for.
So, if you're not generating XML for WCF services, I suggest that you use the XmlSerialiser instead.

Deserialize a soap message

i have one provider extern that send me that xml for test
<?xml version="1.0" encoding="utf-8"?>
<soap:Envelope xmlns:soap="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema">
<soap:Header>
<AVCabeza transactionID="000032" xmlns="http://webservices.patito/Core/">
<Solicitor entityID="WEST" systemType="WEB" />
</AVCabeza>
</soap:Header>
<soap:Body>
<Availability xmlns:a="http://webservices.patito/Availability/"
xmlns:hc="http://webservices.patito/Common/" summaryOnly="true"
xmlns="http://webservices.patito/og/Availability.wsdl">
<a:AvailabilityDetail availReqType="Room">
<a:Estadia>
<hc:StartDate>2009-01-05T00:00:00.0000000-05:00</hc:StartDate>
<hc:EndDate>2009-01-06T00:00:00.0000000-05:00</hc:EndDate>
</a:Estadia>
<a:HotelSearchCriteria>
<a:HotelRef chainCode="WC"/>
</a:HotelSearchCriteria>
</a:AvailabilityDetail>
</Availability>
</soap:Body>
</soap:Envelope>
I want deserializer so i did it
1) i used the xsd for generate c# class
2) Create a new project class library with the class generate.
The structure
WebServicesExterns (Project)
--> Services (Folder)
---> all class
example
namespace WebServicesExterns.Services
<System.CodeDom.Compiler.GeneratedCodeAttribute("wsdl", "2.0.50727.42"), _
System.SerializableAttribute(), _
System.Diagnostics.DebuggerStepThroughAttribute(), _
System.ComponentModel.DesignerCategoryAttribute("code"), _
System.Xml.Serialization.XmlTypeAttribute([Namespace]:="http://webservices.patito/Availability.wsdl")> _
Partial Public Class Availability
'''<comentarios/>
<System.CodeDom.Compiler.GeneratedCodeAttribute("wsdl", "2.0.50727.42"), _
System.SerializableAttribute(), _
System.Diagnostics.DebuggerStepThroughAttribute(), _
System.ComponentModel.DesignerCategoryAttribute("code"), _
System.Xml.Serialization.XmlTypeAttribute([Namespace]:="http://webservices.patito/Core/"), _
System.Xml.Serialization.XmlRootAttribute([Namespace]:="http://webservices.patito/Core/", IsNullable:=false)> _
Partial Public Class AVCabeza
3) After create a test class for try deserializer
using WebServicesExterns.Services;
using System;
using System.Collections;
using System.IO;
using System.Linq;
using System.Runtime.Remoting;
using System.Runtime.Serialization;
using System.Xml;
using System.Xml.Serialization;
using NUnit.Framework;
using System.Reflection;
using System.Runtime.Serialization.Formatters.Soap;
[Test()]
public void ShouldDeserializerSoapMessage()
{
var message = SoapToFromFile(#"C:\rq\Availability.xml");
Assert.IsNotNull(message);
}
public object SoapToFromFile(string filePath)
{
IFormatter formatter;
FileStream fileStream = null;
Object objectFromSoap = null;
try
{
fileStream = new FileStream(filePath, FileMode.Open, FileAccess.Read);
formatter = new SoapFormatter();
objectFromSoap = formatter.Deserialize(fileStream);
}
catch (Exception exception)
{
throw exception;
}
finally
{
if (fileStream != null) fileStream.Close();
}
return objectFromSoap;
}
So, return that error
Parse Error, no assembly associated with Xml key "AVCabeza" "_P1"
Debugging i founded what _P1 is equal to "http://webservices.patito/Core/"
Apparently not found "Type" AVCabeza class
What's wrong?
New update
Look that was the provider give me
one folder with that structure
root
|
-- WSDL_XSD
| |-XSD
| | files with extension .xsd
| |
| --WS
| files with extension .wsdl
|-- XMLSamples
|-files with xml extension that contain soap messsages
well i delete of the xml (previous example) the head
<?xml version="1.0" encoding="utf-8"?>
<Availability xmlns:a="http://webservices.patito/Availability/"
xmlns:hc="http://webservices.patito/Common/" summaryOnly="true"
xmlns="http://webservices.patito/og/Availability.wsdl">
<a:AvailabilityDetail availReqType="Room">
<a:Estadia>
<hc:StartDate>2009-01-05T00:00:00.0000000-05:00</hc:StartDate>
<hc:EndDate>2009-01-06T00:00:00.0000000-05:00</hc:EndDate>
</a:Estadia>
<a:HotelSearchCriteria>
<a:HotelRef chainCode="WC"/>
</a:HotelSearchCriteria>
</a:AvailabilityDetail>
</Availability>
and will try get deserializable availability object but fail this mark
not expected "<Availability .."
now in wsdl_xsd -> ws -> ws i see the availability exists so i think what availability is wrap to availabilitydetail (real request) which is the object i cant delete
availability tag because it have namespace spacefications for childs tags
what think about it?
maybe if i delete availability and insert namespaces (somehow) i could get my de-serializer object

SoapFormatter is not the XML Serializer. You should use the XmlSerializer class.
Also, this is an entire SOAP Message. You'd be better off having them give you the WSDL and using "Add Service Reference". This would give you classes that will do the serialization and deserialization for you.

When you use xsd.exe to generate C# classes from that XML document, in the first step you will get 5 individual .xsd files, providing the
XSD Schema "inferred" from the message. (There are numerous XSD files because you're using numerous XML namespaces in that message)
In particular, XSD.exe will generate code to describe the SOAP Envelope, including the body and header. This is probably not something you want or need to be doing, but the xsd.exe tool infers types for the entire XML document.
Also, the inference engine within xsd.exe is
imprecise. For example, the "StartDate" and "EndDate" child elements of the Estadia element appear to be dates. But xsd.exe won't make that assumption; it will generate an XML schema that marks those things as strings. There are other similar assumptions that xsd.exe makes as it infers. In all cases you probably want to modify the generated xsd files, to match what you truly expect. In the case of StartDate and EndDate, you want to modify the type from xs:string to xs:dateTime.
At that point you can run xsd.exe again, on the .xsd files, using the /c switch, to generate .cs source code. Compile that to get classes that can be used in serialization.
To de-serialize using that generated code, you'd do something like this:
XmlSerializer s1 = new XmlSerializer(typeof(Carlos.Envelope));
Envelope envelope = null;
using(var reader= System.IO.File.OpenText("SampleMessage.xml"))
{
envelope = (Envelope) s1.Deserialize(reader);
}
Then you can open up that Envelope object and get at the various data within it.
Stepping back, you can see that you probably shouldn't be doing this. It's useful and handy to see a sample message, to show you what things ought to look like on the wire. But when generating code to handle classes that serialize to those messages, it's better to start with the source XSD - which is probably available at the service end. This is what John Saunders said in his reply.
If your people can produce a sample message, they probably have the XSD (or the WSDL, which is equivalent) for that message. This would eliminate the requirement for you to infer the xsd, and then modify it to change the broken assumptions, via the inexact process I described above.
The only time you'd need to do that is if you somehow lost the original XSD/WSDL, and you needed to regenerate it.

Excel 2007: XML: choking on missing ss: namespace qualifier

Trying to create an Excel 2007 XML file, using C# and System.XML.Serialization. The root element of an Excel XML file is such:
<Workbook xmlns:html="http://www.w3.org/TR/REC-html40" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:x="urn:schemas-microsoft-com:office:excel" xmlns="urn:schemas-microsoft-com:office:spreadsheet">
Many elements within the default namespace, have attributes qualified with 'ss', which is unnecessary as both 'ss' and the default namespace are both "urn:schemas-microsoft-com:office:spreadsheet".
For example the Style element, appears as:
<Style ss:ID="s21">
<Font x:Family="Swiss" ss:Bold="1"/>
</Style>
When I create my XmlSerializerNamespaces, I add all the namespaces, that Excel wants:
var ns = new XmlSerializerNamespaces();
ns.Add("", "urn:schemas-microsoft-com:office:spreadsheet");
ns.Add("o", "urn:schemas-microsoft-com:office:office");
ns.Add("x", "urn:schemas-microsoft-com:office:excel");
ns.Add("ss", "urn:schemas-microsoft-com:office:spreadsheet");
ns.Add("html", "http://www.w3.org/TR/REC-html40");
In the generated XML, the "" (default) namespace is omitted because it is the same as 'ss'.
My style object is something along the lines of:
[XmlRoot(Namespace = "urn:schemas-microsoft-com:office:spreadsheet")]
public class Styles
{
[XmlAttribute]
public string ID { get; set; }
[XmlAttribute]
public string Name { get; set; }
[XmlElement]
public string Font { get; set; }
}
When I actually generate the XML I am getting:
<ss:Style ID="s21">
<ss:Font x:Family="Swiss" Bold="1"/>
</ss:Style>
I don't think there is anything technically wrong with this. Excel doesn't mind the ss: qualifier on the elements, but it chokes because it can't the 'ID' attribute of Style. It must be hardcoded to look for the literal string 'ss:ID'?
If I omit the 'ss' from the namespaces it produces cleaner XML, but does not make my error go away. I have also tried omitting the 'ss' namespace, and then setting the namespace of the Style object to be literally 'ss', but it generates some temporary namespace named 'd4p1' and makes a real mess.
Any ideas on how to make Excel read valid XML?
And for bonus marks, how can I make the begginning of my XML file look like this:
<?xml version="1.0" encoding="utf-8"?>
<?mso-application progid="Excel.Sheet"?>
Obviously the first part comes with an XML writer, but is there any non-hackish ways to get that second line?
Thanks,
~S

Rather than writing the xml yourself, did you consider using the Open XML SDK 2.0 for Microsoft Office? Besides helping with the manipulation of Open XML docs, it also contains quite a bit of documentation, and seems like it would be well suited for your task:
The Open XML SDK 2.0 for Microsoft Office is built on top of the
System.IO.Packaging API and provides
strongly typed part classes to
manipulate Open XML documents. The SDK
also uses the .NET Framework
Language-Integrated Query (LINQ)
technology to provide strongly typed
object access to the XML content
inside the parts of Open XML
documents.

Trouble with XML Deserialization into XSD generated classes

I have a rather detailed xml file. Below is the top level nodes (I have included the ellipse as the lower level nodes are all well formed and properly filled with data):
<?xml version="1.0" encoding="UTF-8"?>
<config>
<Models>...</Models>
<Data>...</Data>
</config>
I have created an xsd file from using the Visual Studio 2008 command prompt:
xsd sample.xml
This generates the xsd file just fine. I then auto generate classes from the xsd with the command:
xsd sample.xsd /classes
For the deserialization of the xml file into a class object, I'm using the read function in the helper class:
public class XmlSerializerHelper<T>
{
public Type _type;
public XmlSerializerHelper()
{
_type = typeof(T);
}
public void Save(string path, object obj)
{
using (TextWriter textWriter = new StreamWriter(path))
{
XmlSerializer serializer = new XmlSerializer(_type);
serializer.Serialize(textWriter, obj);
}
}
public T Read(string path)
{
T result;
using (TextReader textReader = new StreamReader(path))
{
XmlSerializer deserializer = new XmlSerializer(_type);
result = (T)deserializer.Deserialize(textReader);
}
return result;
}
}
When attempting the deserialization with:
var helper = new XmlSerializerHelper<configModels>();
var obj = new configModels();
obj = helper.Read(filepath);
I receive an error that I have deduced is because the deserializer is looking for the 'Models' node but the corresponding class name was generated as a combination of the root node and the 'Model' node (configModels). Why are the class names generated like this?
I tried to deserialize from the top node using:
var helper = new XmlSerializerHelper<config>();
var obj = new config();
obj = helper.Read(filepath);
Unfortunately, this the results in a slew of errors like the following:
System.InvalidOperationException was unhandled by user code
Message="Unable to generate a temporary class (result=1).
error CS0030: Cannot convert type 'Application.Lease[]' to 'Application.Lease'
error CS0030: Cannot convert type 'Application.CashFlow[]' to 'Application.CashFlow'
...ect.
Can somebody steer me towards what I might be doing wrong with my xsd auto-generating?

XSD.EXE is a good start - but it's far from perfect. Also, based on the XML you provided, XSD.EXE can't always decide for sure whether something is a single instance of an object, or an open-ended array of objects.
This seems to be the case for your two elements - Application.Lease and Application.CashFlow. How are they defined in the generated XSD file? Does that make sense to you? Quite possibly, you'd have to add a little hints, such as:
<xs:element name="Lease" minOccurs="0" maxOccurs="1" />
for an optional property, that's zero or one occurences only. Things like that are really hard for the xsd.exe tool to figure out based on just a single XML sample file.
Marc

Go to your generated class and change all from [][] ---> []

There's an issue with xsd.exe and lists. You have to go into the generated class and manually edit the file to the correct type. I've switched to using Xsd2Code. So far it doesn't seem to have this problem.

Another issue that can cause this problem is that the xml file contents between the tags (meaning the content) is still encoded when it shouldn't be. For example, the <br> tags in my content were still <br> instead of <br />. The xsd generator turned these into elements in the schema then mislabeled them as unbounded since there was more than one found. Unencoding them fixed the problem and generated the classes correctly.

Parse an xml document when namespace is no-longer available

I have a number rather large, complex xml documents that I need to loop through. An xmlns is defined at the top of the document however the url this points to is no longer available.
What's the best way to parse the file to get the important data from it using C#?
I tried to load it into a Dataset but would occasionally receive the errors:
The table (endpoint) cannot be the child table to itself in nested relations.
or
Cannot add a SimpleContent column to a table containing element columns or nested relations.
XPath was my next port of call but I had problems because of the lack of namespace.
I suspect this is seriously limiting my options, but does anyone have any suggestions?
Snippet of the XML document:
<?xml version="1.0" encoding="UTF-8"?>
<cdr:cdr_set xmlns:cdr="http://www.naturalconvergence.com/schema/cdr/v3/cdr">
<!-- Copyright (c) 2001-2009, all rights reserved -->
<cdr:cdr xmlns:cdr="http://www.naturalconvergence.com/schema/cdr/v3/cdr">
<cdr:call_id>2040-1247062136726-5485131</cdr:call_id>
<cdr:cdr_id>1</cdr:cdr_id>
<cdr:status>Normal</cdr:status>
<cdr:responsibility>
<cdr:tenant id="17">
<cdr:name>SpiriTel plc</cdr:name>
</cdr:tenant>
<cdr:site id="45">
<cdr:name>KWS</cdr:name>
<cdr:time_zone>GB</cdr:time_zone>
</cdr:site>
</cdr:responsibility>
<cdr:originator type="sipGateway">
<cdr:sipGateway id="3">
<cdr:name>Audiocodes-91</cdr:name>
</cdr:sipGateway>
</cdr:originator>
<cdr:terminator type="group">
<cdr:group>
<cdr:tenant id="17">
<cdr:name>SpiriTel plc</cdr:name>
</cdr:tenant>
<cdr:type>Broadcast</cdr:type>
<cdr:extension>6024</cdr:extension>
<cdr:name>OLD PMS DDIS DO NOT USE</cdr:name>
</cdr:group>
</cdr:terminator>
<cdr:initiation>Dialed</cdr:initiation>
<cdr:calling_number>02087893850</cdr:calling_number>
<cdr:dialed_number>01942760142</cdr:dialed_number>
<cdr:target>6024</cdr:target>
<cdr:direction>Inbound</cdr:direction>
<cdr:disposition>No Answer</cdr:disposition>
<cdr:timezone>GB</cdr:timezone>
<cdr:origination_timestamp>2009-07-08T15:08:56.727+01:00</cdr:origination_timestamp>
<cdr:release_timestamp>2009-07-08T15:09:26.493+01:00</cdr:release_timestamp>
<cdr:release_cause>Normal Clearing</cdr:release_cause>
<cdr:call_duration>PT29S</cdr:call_duration>
<cdr:redirected>false</cdr:redirected>
<cdr:conference>false</cdr:conference>
<cdr:transferred>false</cdr:transferred>
<cdr:estimated>false</cdr:estimated>
<cdr:interim>false</cdr:interim>
<cdr:segments>
<cdr:segment>
<cdr:originationTimestamp>2009-07-08T15:08:56.727+01:00</cdr:originationTimestamp>
<cdr:initiation>Dialed</cdr:initiation>
<cdr:call_id>2040-1247062136726-5485131</cdr:call_id>
<cdr:originator type="sipGateway">
<cdr:sipGateway id="3">
<cdr:name>Audiocodes-91</cdr:name>
</cdr:sipGateway>
</cdr:originator>
<cdr:termination_attempt>
<cdr:termination_timestamp>2009-07-08T15:08:56.728+01:00</cdr:termination_timestamp>
<cdr:terminator type="group">
<cdr:group>
<cdr:tenant id="17">
<cdr:name>SpiriTel plc</cdr:name>
</cdr:tenant>
<cdr:type>Broadcast</cdr:type>
<cdr:extension>6024</cdr:extension>
<cdr:name>OLD PMS DDIS DO NOT USE</cdr:name>
</cdr:group>
</cdr:terminator>
<cdr:provided_address>01942760142</cdr:provided_address>
<cdr:direction>Inbound</cdr:direction>
<cdr:disposition>No Answer</cdr:disposition>
</cdr:termination_attempt>
</cdr:segment>
</cdr:segments>
</cdr:cdr>
...
</cdr:cdr_set>
Each entry is essentially the same but there are sometimes differences such as some of the fields may be missing, if they aren't required.

These values in an xml file are identifiers, not locators. Unless you are expecting to download a schema, it is not needed at all, and can be "flibble" if needed. I expect the best thing would be to just load it into XmlDocument / XDocument and try to access the data.
For example:
XmlDocument doc = new XmlDocument();
doc.Load("cdr.xml");
XmlNamespaceManager ns = new XmlNamespaceManager(doc.NameTable);
ns.AddNamespace("cdr", "http://www.naturalconvergence.com/schema/cdr/v3/cdr");
XmlElement el = (XmlElement)doc.SelectSingleNode(
"cdr:cdr_set/cdr:cdr/cdr:originator", ns);
Console.WriteLine(el.GetAttribute("type"));
or to loop over the cdr elements:
foreach (XmlElement cdr in doc.SelectNodes("/cdr:cdr_set/cdr:cdr", ns))
{
Console.WriteLine(cdr.SelectSingleNode("cdr:call_id", ns).InnerText);
}
Note that the aliases used in the document are largely unrelated to the aliases used in the XmlNamespaceManager, hence you need to re-declare it. I could have used x as my alias in the C# just as easily.
Of course, if you prefer to work with an object model; run it through xsd (where cdr.xml is your example file):
xsd cdr.xml
xsd cdr.xsd /classes
Now you can load it with XmlSerializer.

alternativley load it into an Xdocument and use linq2XML? ... although you might just get the same error.
I don't know what data you want, so its hard to suggest a query.
I personally prefer the use of XDocument to xmlDocument now in most cases.
the only problem with the automatic generation of an XSD is that it can get your datatypes pretty badly wrong if you are not using a good sized chunk of sample data.