A practical guide to XML and XSD tools available in .NET8 environment, focusing on generating and using C# classes to process some XML valid for some given XSD (technology as of September 2024).
1 Doing XML and XSD related work in .NET8
I was recently doing some work related to XML and XSD processing in .NET8 environment and created several proof-of-concept applications to evaluate the tools available. These articles are the result of my prototyping work.
1.1 List of tools used/tested
Here are the tools used/tested:
- Visual Studio 2022
- XSD.EXE (Microsoft license, part of VS2022)
- XmlSchemaClassGenerator (Open Source/Freeware)
- LinqToXsdCore (Open Source/Freeware)
- Liquid XML Objects (Commercial license)
1.2 Articles in this series
For technical reasons, I will organize this text into several articles:
- XSD Tools in .NET8 – Part1 – VS2022
- XSD Tools in .NET8 – Part2 – C# validation
- XSD Tools in .NET8 – Part3 – XsdExe – Simple
- XSD Tools in .NET8 – Part4 – XsdExe - Advanced
- XSD Tools in .NET8 – Part5 – XmlSchemaClassGenerator – Simple
- XSD Tools in .NET8 – Part6 – XmlSchemaClassGenerator – Advanced
- XSD Tools in .NET8 – Part7 – LinqToXsdCore – Simple
- XSD Tools in .NET8 – Part8 – LinqToXsdCore – Advanced
2 More theory about XML and XSD rules
Here is some more theory about XML and XSD rules.
2.1 Optional Xml-Element and Xml-Attribute
Optional: Does not need to be present in the XML.
For XSD Schema elements:
Optional: minOccurs="0" attribute ->In order to set a schema element as optional, you include the minOccurs="0" attribute
2.2 The Difference Between Optional and Not Required for Xml-Element and Xml-Attribute
Note the difference:
- Optional: Does not need to be present in the XML.
- Not Required: Does not need to have a value.
You can have any combination:
- Optional + Not Required
- Optional + Required
- Not Optional + Not Required
- Not Optional + Required
For XSD Schema elements:
- Optional: minOccurs="0" attribute ->In order to set a schema element as optional, you include the minOccurs="0" attribute
- Required: nillable="true" attribute ->In order to set a schema element as not required, you include the nillable="true" attribute.
String data types are not required by default, though you can force them to be required.
Other data types, such as Boolean, Integer, Date, Time, etc. are all required by default. In order to make one of these data types not required, you must set the nillable attribute equal to true for the element in the schema.
3 Examples of XML and XSD
Here are some sample XML-s and XSD-s I created for test purposes.
3.1 Advanced case
Please note that this example XML/XSD has an Optional and Not-Required Xml-Element. Read the comments inside for more details.
="1.0"="utf-8"
<xs:schema xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
attributeFormDefault="unqualified"
elementFormDefault="qualified"
targetNamespace="https://markpelf.com/BigCompany.xsd"
xmlns:xs="http://www.w3.org/2001/XMLSchema">
<xs:element name="BigCompany">
<xs:complexType>
<xs:sequence>
<xs:element name="CompanyName" type="xs:string" />
<xs:element maxOccurs="unbounded" name="Employee">
<xs:complexType>
<xs:sequence>
<xs:element name="Name_String_NO" type="xs:string" />
<xs:element minOccurs="0" name="City_String_O" type="xs:string" />
</xs:sequence>
</xs:complexType>
</xs:element>
<xs:element maxOccurs="unbounded" name="InfoData">
<xs:complexType>
<xs:sequence>
<xs:element name="Data1_Int_NO_R" type="xs:int" minOccurs="1" nillable="false" />
<xs:element name="Data2_Int_NO_NR" type="xs:int" minOccurs="1" nillable="true" />
<xs:element name="Data3_Int_O_R" type="xs:int" minOccurs="0" nillable="false" />
<xs:element name="Data4_Int_O_NR" type="xs:int" minOccurs="0" nillable="true" />
</xs:sequence>
</xs:complexType>
</xs:element>
</xs:sequence>
</xs:complexType>
</xs:element>
</xs:schema>
="1.0"="utf-8"
<BigCompany xmlns="https://markpelf.com/BigCompany.xsd">
<CompanyName>BigCompanyMMM</CompanyName>
<Employee>
<Name_String_NO>Mark</Name_String_NO>
<City_String_O>Belgrade</City_String_O>
</Employee>
<Employee>
<Name_String_NO>John</Name_String_NO>
</Employee>
<InfoData>
<Data1_Int_NO_R>555</Data1_Int_NO_R>
<Data2_Int_NO_NR>16</Data2_Int_NO_NR>
<Data3_Int_O_R>333</Data3_Int_O_R>
<Data4_Int_O_NR>17</Data4_Int_O_NR>
</InfoData>
<InfoData>
<Data1_Int_NO_R>123</Data1_Int_NO_R>
<Data2_Int_NO_NR xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:nil="true" />
<Data3_Int_O_R>15</Data3_Int_O_R>
<Data4_Int_O_NR xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:nil="true" />
</InfoData>
<InfoData>
<Data1_Int_NO_R>777</Data1_Int_NO_R>
<Data2_Int_NO_NR xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:nil="true" />
</InfoData>
</BigCompany>
4 Using LinqToXsdCore tool to create C# class
We focus in this article on the usage of LinqToXsdCore tool to generate C# class from XSD file.
Here is the tool's basic info.
Tool name: ============================
XSD LinqToXsdCore
License============================
Open Source/Freeware
Where to get it============================
It is NuGet package of type DotnetTool. You need to install it in a special way.
Install Instructions====================================
Create somewhere directory Tool>
I had some additional custom sources. Just using --ignore-failed-sources wasn't enough. I went into Visual Studio -> Tools -> Options -> NuGet Package Manager -> Package Sources and unchecked the extra package sources so only nuget.org and Microsoft Visual Studio Offline Packages was left checked. After doing that I was able to install the tool , and then I re-enabled my custom package sources.
PowerShell, as Administrator
> dotnet new tool-manifest
> dotnet tool install LinqToXsdCore --ignore-failed-sources
You can invoke the tool from this directory using the following commands: 'dotnet tool run LinqToXsd' or 'dotnet LinqToXsd'.
Tool 'linqtoxsdcore' (version '3.4.7') was successfully installed. Entry is added to the manifest file _________\Tool\.config\dotnet-tools.json.
Version============================
Windows PowerShell > ./LinqToXsd.exe
LinqToXsd 3.4.7
Copyright (C) Muhammad Miftah, GitHub Contributors (2019-2022) & Microsoft Corporation (2008-2011)
Help============================
Windows PowerShell > ./LinqToXsd.exe gen
LinqToXsd 3.4.7
Copyright (C) Muhammad Miftah, GitHub Contributors (2019-2022) & Microsoft Corporation (2008-2011)
ERROR(S):
A required value not bound to option name is missing.
-c, --Config (string) Specify the file or folder path to one or more configuration file(s).
Multiple configuration files are merged as one. Incompatible with -AutoConfig
-e, --EnableServiceReference (bool) Imports the 'System.Xml.Serialization' namespace into the generated code.
-a, --AutoConfig (bool) Specify this with a folder containing XSDs and accompanying configuration
files. This argument associate a configuration file with an XSD when one follows the
naming convention: 'schema.xsd' + 'schema.xsd.config' - this is the default convention
used by the 'config -e' verb when you specify a folder. Use this parameter to
associate an XSD with its own configuration settings to prevent those settings being
overriden or merged as the -Config argument would do. Only accepts folder paths.
Incompatible with -Config. Will only generate code for XSDs that have an accompanying
.config file. If no output is generated, run the 'config' verb on the folder first.
-o, --Output (string) Output file name or folder. When specifying multiple input XSDs or input
folders, and this value is a file, all output is merged into a single file. If this
value is a folder, multiple output files are output to this folder.
--help Display this help screen.
--version Display version information.
value pos. 1 Required. (string[]) One or more schema files or folders containing schema files.
Separate multiple files using a comma (,). If folders are given, then the files
referenced in xs:include or xs:import directives are not imported twice. Usage:
'LinqToXsd [verb] <file1.xsd>,<file2.xsd>' or 'LinqToXsd [verb] <folder1>,<folder2>'.
You can also include folder and file paths in the same invocation: 'LinqToXsd [verb]
<file1.xsd>,<folder1>'
====================================================
Usage Examples===================
Instructions to generate C# class
Windows PowerShell > dotnet LinqToXsd gen SmallCompany.xsd
Windows PowerShell > dotnet LinqToXsd gen BigCompany.xsd
====================================
5 Generated C# class
Here is the C# generated by the above tool based on the above presented XSD BigCompany.xsd.
The class's full code produced by this tool is big (835 lines), so I will here show just the beginning of the file.
namespace https.markpelf.com.BigCompany.xsd {
using System;
using System.Collections;
using System.Collections.Generic;
using System.ComponentModel;
using System.IO;
using System.Linq;
using System.Diagnostics;
using System.Xml;
using System.Xml.Schema;
using System.Xml.Linq;
using Xml.Schema.Linq;
public partial class BigCompany : XTypedElement, IXMetaData {
public void Save(string xmlFile) {
XTypedServices.Save(xmlFile, Untyped);
}
public void Save(System.IO.TextWriter tw) {
XTypedServices.Save(tw, Untyped);
}
public void Save(System.Xml.XmlWriter xmlWriter) {
XTypedServices.Save(xmlWriter, Untyped);
}
public static BigCompany Load(string xmlFile) {
return XTypedServices.Load<BigCompany>(xmlFile);
}
public static BigCompany Load(System.IO.TextReader xmlFile) {
return XTypedServices.Load<BigCompany>(xmlFile);
}
public static BigCompany Parse(string xml) {
return XTypedServices.Parse<BigCompany>(xml);
}
public static explicit operator BigCompany(XElement xe) { return XTypedServices.ToXTypedElement<BigCompany>(xe,LinqToXsdTypeManager.Instance as ILinqToXsdTypeManager); }
public override XTypedElement Clone() {
return XTypedServices.CloneXTypedElement<BigCompany>(this);
}
public BigCompany() {
}
[DebuggerBrowsable(DebuggerBrowsableState.Never)]
[EditorBrowsable(EditorBrowsableState.Never)]
protected internal static readonly System.Xml.Linq.XName CompanyNameXName = System.Xml.Linq.XName.Get("CompanyName", "https://markpelf.com/BigCompany.xsd");
Here is the class diagram.
The generated class requires very specific dependencies installed in support.
6 Two C# API styles for Optional Xml-Elements
There are two approaches/styles for marking Optional Xml-Element presence in generated C# code:
- The first is bool_flag_style – using a bool flag to indicate the presence of optional Xml-Element, with flag=false to indicate the Xml-Element was not present.
For example, for some Xml-Element ElemA that, if present, will have value integer, you will get in C# generated two variables “bool ElemA_flag, int ElemA_value”. You need to check if element ElemA was present by first checking the flag ElemA_flag; and then if it is true, you go for the value of ElemA_value. If you do not check flag ElemA_flag first, and just go for the value of ElemA_value you might pick the default int value of zero (0), and you can not know if that is just the default value for C# variable that is always present, but Xml-Element was not present, or that element was present and it actually had the value of zero (0). - The second is nullable_type_style – using a nullable type to indicate the presence of Xml-Element, with value=null to indicate the Xml-Element was not present.
For example, for some Xml-Element ElemA that, if present, will have value integer, you will get in C# generated variable “int? ElemA_nullableValue”. You need to check if element ElemA was present by first checking the ElemA_nullableValue not being null; and then if it is not meaning the element was present, you go for the int value of ElemA_nullableValue.
7 Sample C# app
Here is a sample C# code using the above generated C# class to load and process the above presented XML BigCompanyMMM.xml.
public static void ProcessVer3_Process2(
string? filePath,
Microsoft.Extensions.Logging.ILogger? logger)
{
try
{
logger?.LogInformation(
"+++ProcessVer3_Process2-Start++++++++++++++++++");
logger?.LogInformation("filePath:" + filePath);
TextReader textReader = File.OpenText(filePath ?? String.Empty);
https.markpelf.com.BigCompany.xsd.BigCompany xmlObject =
https.markpelf.com.BigCompany.xsd.BigCompany.Load(textReader);
if (xmlObject != null)
{
logger?.LogInformation("CompanyName:" + xmlObject.CompanyName);
foreach (https.markpelf.com.BigCompany.xsd.BigCompany.EmployeeLocalType item
in xmlObject.Employee)
{
logger?.LogInformation("------------");
logger?.LogInformation("Name_String_NO:" + item.Name_String_NO);
logger?.LogInformation("City_String_O:" + (item.City_String_O ?? "null"));
}
foreach (https.markpelf.com.BigCompany.xsd.BigCompany.InfoDataLocalType item
in xmlObject.InfoData)
{
logger?.LogInformation("------------");
logger?.LogInformation("Data1_Int_NO_R:" + item.Data1_Int_NO_R.ToString());
logger?.LogInformation("Data2_Int_NO_NR:" + (item.Data2_Int_NO_NR?.ToString() ?? "null"));
logger?.LogInformation("Data3_Int_O_R:" + (item.Data3_Int_O_R?.ToString() ?? "null"));
logger?.LogInformation("Data4_Int_O_NR:" + (item.Data4_Int_O_NR?.ToString() ?? "null"));
}
}
else
{
logger?.LogError("xmlObject == null");
}
logger?.LogInformation(
"+++ProcessVer3_Process2-End++++++++++++++++++");
}
catch (Exception ex)
{
string methodName =
$"Type: {System.Reflection.MethodBase.GetCurrentMethod()?.DeclaringType?.FullName}, " +
$"Method: ProcessVer3_Process2; ";
logger?.LogError(ex, methodName);
}
}
And here is the log of execution.
+++ProcessVer3_Process2-Start++++++++++++++++++
filePath:C:\TmpXSD\XsdExample_Ver3\Example01\bin\Debug\net8.0\XmlFiles\BigCompanyMMM.xml
CompanyName:BigCompanyMMM
------------
Name_String_NO:Mark
City_String_O:Belgrade
------------
Name_String_NO:John
City_String_O:null
------------
Data1_Int_NO_R:555
Data2_Int_NO_NR:16
Data3_Int_O_R:333
Data4_Int_O_NR:17
------------
Data1_Int_NO_R:123
Data2_Int_NO_NR:null
Data3_Int_O_R:15
Data4_Int_O_NR:null
------------
Data1_Int_NO_R:777
Data2_Int_NO_NR:null
Data3_Int_O_R:null
Data4_Int_O_NR:null
+++ProcessVer3_Process2-End++++++++++++++++++
8 Analysis
This tool uses “nullable_type_style” approach/style to mark Optional Xml-Element presence in generated C# code. It does not try to indicate a special case of “nill”.
- Data1_Int_NO_R - is int type and always has value
- Data2_Int_NO_NR - is int? type and the meaning is:
a) null – present but “nill” (we have null here even if the element was present but the value was “nill”)
b) int – present and had value - Data3_Int_O_R – is int? type and the meaning is
a) null - means it was not present
b) int – present and had value - Data4_Int_O_NR - is int? type and the meaning is
a) null – means it either was not present or present but “nill”. We can not know which of these 2 cases happened. Strictly speaking, that is a deficiency of this approach/code generated. In some cases, such a distinction might be needed.
b) int – present and had value
It is interesting to look at this tool regarding the ability in the case of Xml-Element Data4_Int_O_NR to indicate all three states: “not-present”, “present-nill”, “present-int”. I do not see that using the modern API style/approach “nullable_type_style” the that can be achieved. If Data4_Int_O_NR is null, we can not know if it was “not-present” or “present-nill”.
9 Conclusion
This tool LinqToXsdCore is very interesting and available as freeware. Code C# class generated was very big (it includes some extra methods like Load, Save, Clone), but worked solid in my test. The issue is only that “nillable” option is not handled well at all. Theoretically, that is a problem and for some users, it can be important, if they want a tool that follows all XML specifications.
It can be of interest to users who want to use nullable_type_style API, which is generally a more modern approach to handling the Optional Xml-Elements.
Some things make me a bit reserved about using this tool, like why the generated class is so big. Even more than that is the question of those very specific dependencies libraries, are they maintained, and reliable, and will they cause any conflicts with other XML libraries? Answers to these questions would require more studying of this tool before deciding whether to use it in production.
10 References
[1] XML Schema
https://www.w3schools.com/xml/xml_schema.asp
