Suppose I have this class:
public class GroceryListItem()
{
public GroceryList { get; private set; }
public GroceryListItem(GroceryList groceryList)
{
GroceryList = groceryList;
}
}
What is the NHibernate mapping file access strategy for this scenario? (i.e. <one-to-many name="GroceryList" column="XXX" access="?????" />)
It turns out the answer is pretty simple -- no special access is required. NHibernate is smart enough to work this out on its own. In other words, the code in my question works correctly with the following line in the mapping file:
<one-to-many name="GroceryList" column="XXX" />
Use access="readonly" in the newer versions, or create your own PropertyAccessor or use any of the other approaches described here:
http://blog.schuager.com/2008/12/nhibernate-read-only-property-access.html
Related
I have a class that I need to map that looks a bit like this
public class Foo
{
public string Name { get; set; }
// other stuff
public IDictionary<Bar, decimal> Bars { get; set; }
}
my question is how can I map this with NHibernate, I was thinking I can treat the dictionary as a HasMany, since I can recreate a Bar from a string, however it feels a bit crude and I wonder if there is a better way
cheers
<map name="Bars">
<key column="..." />
<map-key-many-to-many class="Bar" />
<element type="String" /><!--or decimal, or anything else-->
</map>
It depends what the decimal might be, but NHibernate does provide a <map> collection type specifically for Dictionary style mappings.
may be this would help:
nhibernate-mapping-map
Does anybody know if it is possible to control the names of the types generated through Castle DynamicProxy? I was hoping to take advantage of the ability to persist the assembly generated by Castle to add some additional classes with some specific functionality to my project, but I would like to be able to control the names of these generated proxy types. Any help would be greatly appreciated.
I actually plan to persist instances of these classes as well as instances of the original classes that are the sources of the proxies with NHibernate. So, I need these names to be consistent across multiple generations of the assembly.
I did some interesting digging. Specifying proxy names appears to be possible using an INamingScope, but it is far from straightforward to get the INamingScope wedged in. You would need to create your own ProxyFactoryFactory, which would create a ProxyFactory identical to NHibernate.ByteCode.Castle.ProxyFactory, except it would initilize ProxyGenerator:
public class CustomProxyFactory : AbstractProxyFactory {
private static readonly ProxyGenerator ProxyGenerator = new ProxyGenerator(new CustomProxyBuilder());
// remainder of code is identical
}
public class CustomProxyBuilder : DefaultProxyBuilder {
public CustomProxyBuilder() : base(new CustomModuleScope()) {}
}
public class CustomModuleScope : ModuleScope {
public CustomModuleScope() : base(false, false, new CustomNamingScope(), DEFAULT_ASSEMBLY_NAME, DEFAULT_FILE_NAME, DEFAULT_ASSEMBLY_NAME, DEFAULT_FILE_NAME) {}
}
public class CustomNamingScope : INamingScope {
public CustomNamingScope() {}
private CustomNamingScope(INamingScope parent) {
ParentScope = parent;
}
public string GetUniqueName(string suggestedName) {
// your naming logic goes here
}
public INamingScope SafeSubScope() {
return new CustomModuleScope(this);
}
public INamingScope ParentScope { get; private set; }
}
I honestly haven't tried running or compiling any of this. Just digging through the NHibernate and Castle.Core source code. Hopefully it gives you some ideas...
Take a look at the ProxyGenerators project in NHContrib. It allows you to pre-generate NHibernate's lazy loading proxies.
http://nhforge.org/wikis/proxygenerators10/default.aspx
Whether you use the ProxyGenerators or not, you integrate your custom proxies into NHibernate via the Proxy Factory Factory. In hibernate.cfg.xml:
<hibernate-configuration xmlns="urn:nhibernate-configuration-2.2">
<session-factory>
<property name="proxyfactory.factory_class">YOUR_PROXY_FACTORY_FACTORY</property>
</session-factory>
</hibernate-configuration>
I just had a NHibernate related problem where I forgot to map one property of a class.
A very simplified example:
public class MyClass
{
public virtual int ID { get; set; }
public virtual string SomeText { get; set; }
public virtual int SomeNumber { get; set; }
}
...and the mapping file:
<?xml version="1.0" encoding="utf-8" ?>
<hibernate-mapping xmlns="urn:nhibernate-mapping-2.2"
assembly="MyAssembly"
namespace="MyAssembly.MyNamespace">
<class name="MyClass" table="SomeTable">
<property name="ID" />
<property name="SomeText" />
</class>
</hibernate-mapping>
In this simple example, you can see the problem at once:
there is a property named "SomeNumber" in the class, but not in the mapping file.
So NHibernate will not map it and it will always be zero.
The real class had a lot more properties, so the problem was not as easy to see and it took me quite some time to figure out why SomeNumber always returned zero even though I was 100% sure that the value in the database was != zero.
So, here is my question:
Is there some simple way to find this out via NHibernate?
Like a compiler warning when a class is mapped, but some of its properties are not.
Or some query that I can run that shows me unmapped properties in mapped classes...you get the idea.
(Plus, it would be nice if I could exclude some legacy columns that I really don't want mapped.)
EDIT:
Okay, I looked at everything you proposed and decided to go with the meta-data API...that looks the easiest to understand for me.
Now that I know what to search for, I found some examples which helped me to get started.
So far, I have this:
Type type = typeof(MyClass);
IClassMetadata meta = MySessionFactory.GetClassMetadata(type);
PropertyInfo[] infos = type.GetProperties();
foreach (PropertyInfo info in infos)
{
if (meta.PropertyNames.Contains(info.Name))
{
Console.WriteLine("{0} is mapped!", info.Name);
}
else
{
Console.WriteLine("{0} is not mapped!", info.Name);
}
}
It nearly works, except one thing:
IClassMetadata.PropertyNames returns the names of all the properties except the ID.
To get the ID, I have to use IClassMetadata.IdentifierPropertyName.
Yes, I could save .PropertyNames in a new array, add .IdentifierPropertyName to it and search that array.
But this looks strange to me.
Is there no better way to get all mapped properties including the ID?
You could use the NHibernate meta-data API to find the mapped properties, and reflection to find all the properties.
Edit No, there isn't any other way list all the properties including the id. It isn't that hard to use:
foreach (PropertyInfo info in infos)
{
if (meta.PropertyNames.Contains(info.Name) || info.Name = meta.IdentifierPropertyName)
{
Console.WriteLine("{0} is mapped!", info.Name);
}
else
{
Console.WriteLine("{0} is not mapped!", info.Name);
}
}
There are two tools I'm aware of that can help with this:
Fluent NHibernate persistence specification testing
Nhibernate Ghostbuster
but they don't specifically address the problem you had with an unmapped property. The best solution is to write good unit tests that ensure that the properties you want to persist are persisted correctly. It's tedious but necessary.
nHibernate is giving the error : Custom type does not implement UserCollectionType: myApp.Domain.OrderLineCollection.
BindingList implements IList, so why is nHibernate trying to use UserCollectionType instead of IList?
public class OrderHeader
{
public virtual int OrderHeaderId { get; set; }
public virtual string OrderNumber { get; set; }
public virtual OrderLineCollection Line { get; set; }
}
public class OrderLineCollection : BindingList<OrderHeader> { }
public class OrderHeaderMap : ClassMap<OrderHeader>
{
public OrderHeaderMap()
{
WithTable("Orders");
Id(x => x.OrderHeaderId, "OrderId").GeneratedBy.Identity();
Map(x => x.OrderNumber);
HasMany(x => x.Line).WithKeyColumn("OrderHeaderId").AsList();
}
}
<list name="Line">
<key column="OrderHeaderId" />
<index />
<one-to-many class="myApp.Domain.OrderLine, myApp.Domain, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null" />
</list>
NHibernate has it's own custom typed list which implements IList underneath.
I'm afraid you won't be able to use yours without creating nHibernate UserType.
But i might be wrong and would be glad to hear why. :)
You might want to check the XML that's created by fluentNHibernate - it's quite possible they take the type of the Line property and set it explicitly.
This should work if you don't set the type explicitly. I tried implementing a custom collection deriving from IList - and it worked when I didn't specify the type on the bag/list whatever in the mapping.
Ok, I did a quick test Arnis L. is right - it probably won't work without implementing UserCollectionType. In my experience, it's a pain to implement .
(somehow I remembered doing something like this but I guess my mind's playing tricks on me)
I look at the NHibernate source code and at least for PersistentBag and PersistentList NHibernate will instanciate a ArrayList object as the back end list, not a OrderLineCollection as one could thought. When you implement IUserColletionType there is a method who tells NHibernate what collection it should create, and also what Persistent collection Hibernate should use to sav. Take a look at this link might help a lot. But I still cant do Nhibernate work with BindingList.
I've seen several questions related to properly mapping an enum type using NHibernate.
This article by Jeff Palermo showed me how to do that properly by creating a custom type. I use Schema Export to create my DB during my dev cycles, but this method breaks my export statement. Is there a way to specify the type of the column on export?
Here is my enum code:
public enum OperatorCode
{
CodeA,
CodeB,
CodeC,
CodeD
}
Here is my custom type:
public class OperatorCodeType:EnumStringType
{
public OperatorCodeType():base(typeof(OperatorCode),20)
{
}
}
Here is my property in my mapping file:
<property name="OperatorCode" column="OperatorCode" type="OperatorCodeType" />
And finally here is my class declaration for that property:
public virtual OperatorCode OperatorCode { get; set; }
Is it even possible to do this?
I have not tested it, but you can use the Column declaration within a property to specify the sql type. Example from the docs:
<property name="Foo" type="String">
<column name="foo" length="64" not-null="true" sql-type="text"/>
</property>
Granted this is a string, but you may want to try it with the type of OperatorCodeType, column sql-type as text or nvarchar or whatever works.
If you try it, let me know? Not near my dev machine at the moment.