Lesser-Known NHibernate Features: Versioning and Optimistic Concurrency

NHibernate supports the notion of entity version. An entity version is the value of some entity property that is mapped as versioned. There are several strategies for versioning:

  • A numeric counter;
  • The current (UTC) timestamp;
  • The database timestamp;
  • The system elapsed ticks (DateTime.Ticks);
  • A database-specific row version (ROWVERSION in SQL Server, ORA_ROWSCN in Oracle, etc).


We map a version property as this:

public class VersionedEntity


    public int Id { get; protected set; }

    public int Version { get; protected set; }


public class VersionedEntityClassMapping<VersionedEntity>


    public VersionedEntityClassMapping()


        this.Id(x => x.Id, x => x.Generator(Generators.Identity));

        this.Version(x => x.Version, x => x.Type(NHibernateUtil.Int32));




If you wish to use a strategy other than auto-incrementing versioning, just replace NHibernateUtil.Int32 by another instance of a class implementing IVersionType, such as NHibernateUtil.Ticks, NHibernateUtil.Timestamp, NHibernateUtil.UtcDateTime, NHibernateUtil.DbTimestamp, NHibernateUtil.Binary (for SQL Server), or a similar one – or roll out your own strategy.

Close to the notion of versioning comes optimistic concurrency. Using optimistic concurrency control, when a record is about to be updated, it is checked against a pre-saved value or set of values (the version), together with its primary key. If it doesn’t match, the number of affected records will be 0 instead of 1, and we know something went wrong: because the primary key doesn’t change, then it has to be the version:

UPDATE some_table

SET some_column = @p1

WHERE id = @p2

AND version = @p3

Optimistic concurrency can use:

  • The version property (if any);
  • All of the entity’s properties;
  • All the dirty properties.

We just need to tell NHibernate which strategy we want to use:

public class VersionedEntityClassMapping<VersionedEntity>


    public VersionedEntityClassMapping()


        this.Id(x => x.Id, x => x.Generator(Generators.Identity));

        this.Version(x => x.Version, x => x.Type(NHibernateUtil.Int32));

        this.OptimisticLock(OptimisticLockMode.Version);//NHibernate 4.0, no need to specify since this is the default




And that’s it. Happy versioning!

Lesser-Known NHibernate Features: Filtering an Unloaded Collection

Suppose you have an entity with a collection of other entities (one to many, many to many); this collection is normally represented by some property implementing IEnumerable<T>, or one some more specific interface.

If we want to query this collection, we normally have to load all of its items. For indexed collections, it is possible to use extra-lazy laziness, which means NHibernate will only load one item at a time, instead of all at once, but this might even be worse, because it would result in multiple SELECT statements.

NHibernate offers a nice mechanism to filter and order a collection without actually loading all of its items; it’s the CreateFilter method of the ISession, and we use it like this:

   1: //get an order

   2: var order = session.Query<Order>().First();


   4: //all order details of this order where quantity is greater than one

   5: var detailsWithMoreThanOneItem = session.CreateFilter(order.Details, "WHERE Quantity > 1").List<OrderDetail>();


   7: //all order details of this order sorted by their quantity in descending order

   8: var detailsSortedByQuantityInDescendingOrder = session.CreateFilter(o.Details, "ORDER BY Quantity DESC").List<OrderDetail>();

I even once wrote a helper method for querying collections with LINQ without actually loading them; its usage is:

   1: //all order details of this order where quantity is greater than one

   2: var detailsWithMoreThanOneItem = order.Details.Query().Where(d => d.Quantity > 1).ToList();

The Query extension method will check to see if the collection has already been loaded, in which case, it resorts to LINQ to Objects, otherwise, it generates a proper LINQ query. You can find the source code here.

Entity Framework Pitfalls: Command Interceptors and Identity Keys

The IDbCommandInterceptor (sorry, no official documentation) interface was introduced in Entity Framework 6 as part of the new interception and logging API, and allows the interception of the SQL and its parameters that are sent to the database as the result of DbContext CRUD operations.

It offers two methods for each of the basic ADO.NET operations, those defined in DbCommand, one called before the operation is executed, and the other called afterwards:

  • ExecuteNonQuery (UPDATEs, INSERTs, DELETEs) –> NonQueryExecuting, NonQueryExecuted;
  • ExecuteScalar (SELECTs returning a single value) –> ScalarExecuting, ScalarExecuted;
  • ExecuteReader (SELECTs returning rows) –> ReaderExecuting, ReaderExecuted.

As usual, the methods ending with “ing” are executed before and those ending with “ed” are executed afterwards, always synchronously.

One might naively assume that INSERTs would always trigger a NonQueryExecuting/NonQueryExecuted call, and indeed it is so, unless we have an IDENTITY primary key, in which case, Entity Framework will instead call ReaderExecuting/ReaderExecuted. It’s easy to understand why: when we use an IDENTITY, we need to retrieve the generated value immediately after the INSERT, hence Entity Framework will generate code like:

   1: INSERT INTO [dbo].[SomeTable] (...)

   2: VALUES (...)

   3: SELECT [Id]

   4: FROM [dbo].[SomeTable]


The INSERT and SELECT are combined in the same command, which justifies the choice of ReaderExecuting/ReaderExecuted. Because in Entity Framework we always use IDENTITY – don’t we? –, it’s ReaderExecuting/ReaderExecuted that you should be implementing if you want to change the INSERT SQL or its parameters.

Custom Entity Framework Code First Convention for Discriminator Values

Since version 6, Entity Framework Code First allows the injection of custom conventions. These conventions define rules that will be applied by default to all mapped entities and properties, unless explicitly changed.

The conventions API includes a couple of interfaces: IConvention (marker only, should always be included), IConceptualModelConvention<T> (for the conceptual space of the model) and IStoreModelConvention<T> (for the store, or physical, side of the model). Worthy of mention, there is also a convenience class, Convention, that allows access to all mapped types and properties and doesn’t override any of the other conventions, and also TypeAttributeConfigurationConvention<T>, for tying a convention to a custom attribute. Some of the included attributes leverage these interfaces to configure some aspects of the mappings at design time, other configuration needs to be done explicitly in an override of OnModelCreating.

Entity Framework permits using a column for distinguishing between different types, when the Table Per Class Hierarchy / Single Table Inheritance pattern (please see Entity Framework Code First Inheritance for more information) is used for mapping a hierarchy of classes to a single table, as part of “soft delete” solutions, or, less known, for differentiating between multiple tenants. This column is called a discriminator.

In order to configure an entity to use a discriminator column, there is no out of the box attribute, so we must resort to code configuration:

   1: protected override void OnModelCreating(DbModelBuilder modelBuilder)

   2: {

   3:     modelBuilder.Entity<MyMultiTenantEntity>().Map(m => m.Requires("tenant_id").HasValue("first_tenant"));


   5:     base.OnModelCreating(modelBuilder);

   6: }

Because there’s really no need to keep repeating this code, let’s implement an attribute for indicating a discriminator column in an entity:

   1: [Serializable]

   2: [AttributeUsage(AttributeTargets.Class, AllowMultiple = false, Inherited = true)]

   3: public sealed class DiscriminatorAttribute : Attribute

   4: {

   5:     public DiscriminatorAttribute(String columnName, Object discriminatorValue)

   6:     {

   7:         this.ColumnName = columnName;

   8:         this.DiscriminatorValue = discriminatorValue;

   9:     }


  11:     public String ColumnName { get; private set; }


  13:     public Object DiscriminatorValue { get; private set; }


  15:     public override Boolean Equals(Object obj)

  16:     {

  17:         var other = obj as DiscriminatorAttribute;


  19:         if (other == null)

  20:         {

  21:             return (false);

  22:         }


  24:         return ((this.ColumnName == other.ColumnName) && (Object.Equals(this.DiscriminatorValue, other.DiscriminatorValue) == true));

  25:     }


  27:     public override Int32 GetHashCode()

  28:     {

  29:         return (String.Concat(this.ColumnName, ":", this.DiscriminatorValue).GetHashCode());

  30:     }

  31: }

As you can see, the DiscriminatorAttribute attribute can only be applied to a class, at most once. This makes sense, because most likely you will only have a single discriminator column per entity:

   1: [Discriminator("tenant_id", "first_tenant")]

   2: public class MyMultiTenantEntity

   3: {

   4:     //...

   5: }

You need to specify both a column name and a discriminator value, which can be of any type, usually, a string or an integer.

Now, let’s write a custom convention that knows how to handle our custom attribute and perform the mapping:



   1: public sealed class DiscriminatorConvention : TypeAttributeConfigurationConvention<DiscriminatorAttribute>

   2: {

   3:     private static readonly MethodInfo entityMethod = typeof(DbModelBuilder).GetMethod("Entity");

   4:     private static readonly MethodInfo hasValueMethod = typeof(ValueConditionConfiguration).GetMethods().Single(m => (m.Name == "HasValue") && (m.IsGenericMethod == false));


   6:     private readonly DbModelBuilder modelBuilder;

   7:     private readonly ISet<Type> types = new HashSet<Type>();


   9:     public DiscriminatorConvention(DbModelBuilder modelBuilder)

  10:     {

  11:         this.modelBuilder = modelBuilder;

  12:     }


  14:     public override void Apply(ConventionTypeConfiguration configuration, DiscriminatorAttribute attribute)

  15:     {

  16:         if (this.types.Contains(configuration.ClrType) == true)

  17:         {

  18:             //if the type has already been processed, bail out

  19:             return;

  20:         }


  22:         //add the type to the list of processed types

  23:         this.types.Add(configuration.ClrType);


  25:         dynamic entity = entityMethod.MakeGenericMethod(configuration.ClrType).Invoke(modelBuilder, null);


  27:         Action<dynamic> action = arg =>

  28:         {

  29:             var valueConditionConfiguration = arg.Requires(attribute.ColumnName);

  30:             hasValueMethod.Invoke(valueConditionConfiguration, new Object[] { attribute.DiscriminatorValue });

  31:         };


  33:         entity.Map(action);

  34:     }

  35: }

This class uses a bit of dynamics and reflection because types are not known at compile time, and hence we cannot use generics directly. Because the Apply method will be called multiple times, we need to keep track of which entities have already been processed by this convention, so as to avoid reprocessing them. We need to pass it the instance of DbModelBuilder, because otherwise our custom convention would have no way to apply the mapping, but I think it is a reasonable trade off.

Et voilà! In order to make use of it, we need to register the convention in OnModelCreating:

   1: protected override void OnModelCreating(DbModelBuilder modelBuilder)

   2: {

   3:     modelBuilder.Conventions.Add(new DiscriminatorConvention(modelBuilder));


   5:     base.OnModelCreating(modelBuilder);

   6: }

And that’s it! Happy conventions! Winking smile

Entity Framework Pitfalls: Mapping Discriminator Columns

When you use a discriminator column, that is, a column that holds a value that tells Entity Framework what type the row refers to, or what restriction it shall use when querying the entity, you cannot map this column as a property in your entity.

For example, imagine you want to use “soft deletes”, that is, use a database column to represent the “deleted” state of a record instead of actually physically deleting it from the table; you might have an IS_DELETED column of some integer type, that would hold either a 1 or a 0, depending on whether the record is deleted (1) or not (0). In that case, you would build a mapping like this in OnModelCreating:

   1: modelBuilder.Entity<MySoftDeletableEntity>().Map<MySoftDeletableEntity>(m => m.Requires("is_deleted").HasValue(0));

However, you won’t be able to have a corresponding IsDeleted property in the MySoftDeletableEntity, because Entity Framework will throw an exception complaining about it. It is somewhat sad, because it might be useful to refer to it, but that’s the way it is.

Lesser-Known NHibernate Features: Result Transformers

A result transformer, in NHibernate, is some class that implements the IResultTransformer interface:

   1: public interface IResultTransformer

   2: {

   3:     IList TransformList(IList collection);

   4:     Object TransformTuple(Object[] tuple, String[] aliases);

   5: }

Most query APIs, except LINQ, support specifying a result transformer. So, what is a result transformer used for? Just what the name says: it turns the values obtained from a query into some object. Normally, we just let NHibernate transform these values into instances of our entities, but we may want to do something different, either because we haven’t mapped some class that we want to use, or because we are not returning all of the entity’s properties, etc.

NHibernate includes some result transformers:

  • AliasToBeanResultTransformer: allows to transform a result to a user specified class which will be populated via setter methods or fields matching the alias names;
  • AliasToBeanConstructorResultTransformer: identical to AliasToBeanResultTransformer, but we specify a constructor for creating new instances of the target class;
  • AliasToEntityMapResultTransformer: returns a dictionary where the keys are the aliases and the values the corresponding columns;
  • AliasedTupleSubsetResultTransformer: ignores a tuple element if its corresponding alias is null;
  • CacheableResultTransformer: used to transform tuples to a value(s) that can be cached;
  • DistinctRootEntityResultTransformer: for joined queries, returns distinct root entities only;
  • PassThroughResultTransformer: just returns the row as it was obtained from the database;
  • RootEntityResultTransformer; returns the root entity of a joined query;
  • ToListResultTransformer: transforms each result row from a tuple into a IList, such that what you end up with is a IList of ILists.

All of these can be obtained from static properties in class NHibernate.Transform.Transformers. NHibernate implicitly uses some of these, for example, LINQ queries always use DistinctRootEntityResultTransformer.

It is easy to build our own transformer. Have a look at the following example:

   1: public class ExpressionsResultTransformer : IResultTransformer

   2: {

   3:     private readonly Type type;

   4:     private readonly Func<Object> constructorFunc;

   5:     private readonly ConstructorInfo constructor;

   6:     private readonly Object[] parameters;

   7:     private readonly Dictionary<Int32, String> expressions = new Dictionary<Int32, String>();

   8:     private PropertyDescriptorCollection props;


  10:     public ExpressionsResultTransformer(ConstructorInfo constructor, params Object[] parameters)

  11:     {

  12:         this.constructor = constructor;

  13:         this.parameters = parameters;

  14:     }


  16:     public ExpressionsResultTransformer(Func<Object> constructorFunc)

  17:     {

  18:         this.constructorFunc = constructorFunc;

  19:     }


  21:     public ExpressionsResultTransformer(Type type)

  22:     {

  23:         this.type = type;

  24:     }


  26:     private Object CreateInstance()

  27:     {

  28:         if (this.type != null)

  29:         {

  30:             return (Activator.CreateInstance(this.type));

  31:         }

  32:         else if (this.constructorFunc != null)

  33:         {

  34:             return (this.constructorFunc());

  35:         }

  36:         else

  37:         {

  38:             return (this.constructor.Invoke(this.parameters));

  39:         }

  40:     }


  42:     public ExpressionsResultTransformer Add(Int32 index, String property)

  43:     {

  44:         this.expressions[index] = property;

  45:         return (this);

  46:     }


  48:     public ExpressionsResultTransformer Add(params String[] properties)

  49:     {

  50:         foreach (var property in properties)

  51:         {

  52:             this.Add(property);

  53:         }


  55:         return (this);

  56:     }


  58:     public ExpressionsResultTransformer Add(String property)

  59:     {

  60:         var max = this.expressions.Keys.DefaultIfEmpty(-1).Max();

  61:         return (this.Add(max + 1, property));

  62:     }


  64:     public static ExpressionsResultTransformer Add<T>(params String[] properties) where T : new()

  65:     {

  66:         return (new ExpressionsResultTransformer<T>().Add(properties));

  67:     }


  69:     #region IResultTransformer Members


  71:     IList IResultTransformer.TransformList(IList collection)

  72:     {

  73:         return (collection);

  74:     }


  76:     Object IResultTransformer.TransformTuple(Object[] tuple, String[] aliases)

  77:     {

  78:         var entity = this.CreateInstance();


  80:         if (this.props == null)

  81:         {

  82:             this.props = TypeDescriptor.GetProperties(entity);

  83:         }


  85:         foreach (var expression in this.expressions)

  86:         {

  87:             this.props[expression.Value].SetValue(entity, tuple[expression.Key]);

  88:         }


  90:         return (entity);

  91:     }


  93:     #endregion

  94: }


  96: public sealed class ExpressionsResultTransformer<T> : ExpressionsResultTransformer where T : new()

  97: {

  98:     public ExpressionsResultTransformer() : base(typeof(T))

  99:     {

 100:     }


 102:     public ExpressionsResultTransformer(Func<T> constructor) : base(new Func<Object>(() => (Object) constructor()))

 103:     {

 104:     }


 106:     public ExpressionsResultTransformer(ConstructorInfo constructor, params Object[] parameters) : base(constructor, parameters)

 107:     {

 108:     }

 109: }

The TransformTuple method is the one used to turn each returned record into an instance of something. TransformList is called at the end, when all the records have been processed.

The ExpressionResultTransformer class allows us to select which indexes, in the database record, map to which properties in some entity. For our convenience, it offers a number of options to construct an instance (type, constructor + parameters and delegate). We would use it like this:

   1: var products = session.CreateSQLQuery("select p.Name, p.Price, p.ProductId from Product p").SetResultTransformer(ExpressionsResultTransformer.Add<Product>("Name", "Price", "ProductId")).List<Product>();

Getting the SQL for HQL and Criteria Queries

OK, so, I already showed how we can get the SQL that was generated from a LINQ query. Of course, we can do the same for both HQL and Criteria APIs as well (QueryOver is just a wrapper around Criteria, mind you).

So, for HQL (and SQL), it goes like this:

   1: public static class QueryExtensions

   2: {

   3:     private static readonly PropertyInfo sessionProperty = typeof (QueryImpl).GetProperty("Session", BindingFlags.NonPublic | BindingFlags.Instance);


   5:     public static String ToSql(this IQuery query)

   6:     {

   7:         if (query is QueryImpl)

   8:         {

   9:             var session = sessionProperty.GetValue(query, null) as ISession;

  10:             var sessionImpl = session.GetSessionImplementation();

  11:             var translatorFactory = new ASTQueryTranslatorFactory();

  12:             var translators = translatorFactory.CreateQueryTranslators(query.QueryString, null, false, sessionImpl.EnabledFilters, sessionImpl.Factory);


  14:             return translators[0].SQLString;

  15:         }

  16:         else if (query is SqlQueryImpl)

  17:         {

  18:             return (query.QueryString);

  19:         }


  21:         throw (new NotSupportedException("Query type is not supported."));

  22:     }

  23: }

You can pass any implementation if IQuery, such as one produced from ISession.CreateQuery() or ISession.CreateSQLQuery(). The static field is merely for performance reasons.

As for Criteria:

   1: public static class CriteriaExtensions

   2: {

   3:     public static String ToSql(this ICriteria criteria)

   4:     {

   5:         var criteriaImpl = criteria as CriteriaImpl;

   6:         var sessionImpl = criteriaImpl.Session;

   7:         var factory = sessionImpl.Factory;

   8:         var implementors = factory.GetImplementors(criteriaImpl.EntityOrClassName);

   9:         var loader = new CriteriaLoader(factory.GetEntityPersister(implementors[0]) as IOuterJoinLoadable, factory, criteriaImpl, implementors[0], sessionImpl.EnabledFilters);


  11:         return loader.SqlString.ToString();

  12:     }

  13: }

And finally, QueryOver, just a small wrapper around the Criteria version:

   1: public static class QueryOverExtensions

   2: {

   3:     public static String ToSql(this IQueryOver queryOver)

   4:     {

   5:         var criteria = queryOver.UnderlyingCriteria;

   6:         return (criteria.ToSql());

   7:     }

   8: }

Hope you find this useful! Winking smile

Adding Support for ISupportInitialize in NHibernate

The .NET ISupportInitialize interface is used when we want to support staged initialization for objects. Its BeginInit method is called when initialization is about to start and EndInit when it is finished.

If we want, it is easy to add support for it in NHibernate. An option would be:

  • BeginInit is called when the object is instantiated, like when NHibernate has loaded a record from the database and is about to hydrate the entity, and immediately after the Id property is set;
  • EndInit is called after all properties are set.

We do this by using a custom interceptor, like we have in the past. We start by writing a class that inherits from EmptyInterceptor, and implements the listener interface for the PostLoad event, IPostLoadEventListener:

   1: public sealed class SupportInitializeInterceptor : EmptyInterceptor, IPostLoadEventListener

   2: {

   3:     private ISession session = null;


   5:     public override void SetSession(ISession session)

   6:     {

   7:         this.session = session;

   8:         base.SetSession(session);

   9:     }


  11:     public override Object Instantiate(String clazz, EntityMode entityMode, Object id)

  12:     {

  13:         var listeners = (this.session.SessionFactory as SessionFactoryImpl).EventListeners;

  14:         var metadata = this.session.SessionFactory.GetClassMetadata(clazz);

  15:         var proxy = metadata.Instantiate(id, entityMode);

  16:         var initializable = proxy as ISupportInitialize;


  18:         if (initializable != null)

  19:         {

  20:             initializable.BeginInit();

  21:         }


  23:         if (listeners.PostLoadEventListeners.OfType<SupportInitializeInterceptor>().Any() == false)

  24:         {

  25:             listeners.PostLoadEventListeners = listeners.PostLoadEventListeners.Concat(new IPostLoadEventListener[] { this }).ToArray();

  26:         }


  28:         return (proxy);

  29:     }


  31:     #region IPostLoadEventListener Members


  33:     void IPostLoadEventListener.OnPostLoad(PostLoadEvent @event)

  34:     {

  35:         var initializable = @event.Entity as ISupportInitialize;


  37:         if (initializable != null)

  38:         {

  39:             initializable.EndInit();

  40:         }

  41:     }


  43:     #endregion

  44: }

Then, before creating a session factory, we need to register it in the Configuration instance:

   1: var sessionFactory = cfg.SetInterceptor(new SupportInitializeInterceptor()).BuildSessionFactory();

Now, if your entity implements ISupportInitialize, NHibernate will automagically call its methods at the proper time. As simple as this! Winking smile

Lesser-Known NHibernate Features – Generating Database Scripts

As you may know, NHibernate knows how to generate the database for you from its model. Any of the supported databases will do.

What you may not know is that you can simply generate the SQL scripts that would be used to either generate the database from scratch or just update it to match the model as it is. For that you use the SchemaExport and SchemaValidator classes.

SchemaValidator takes a Configuration instance and basically throws an exception if, when its Validate method is called, the database does not match the model.

SchemaExport has far more options; it can:

  • Create or drop a database model based on the entity model;
  • Output the SQL to a file, either executing it at the same time or not;
  • Execute the SQL to another arbitrary database connection;
  • Execute a custom action after each line of the generated SQL script.

An example:

   1: var validator = new SchemaValidator(cfg);


   3: try

   4: {

   5:     validator.Validate();

   6: }

   7: catch

   8: {

   9:     var export = new SchemaExport(cfg).SetOutputFile("Script.sql");


  11:     //send the script to standard output and execute it

  12:     export.Execute(useStdOut: true, execute: true, justDrop: false);

  13:     //or

  14:     //send all lines of the script to the System.Console.WriteLine method

  15:     export.Execute(scriptAction: System.Console.WriteLine, execute: false, justDrop: false);

  16: }

NHibernate Pitfalls: Versioned Entities Are Not Batcheable

This is part of a series of posts about NHibernate Pitfalls. See the entire collection here.

This is a problem similar to the one with native id generators. Basically, because there are different versioning strategies – timestamp, counter, native, etc, – NHibernate needs to issue a SELECT after a versioned entity is INSERTEd or UPDATEd. While this wouldn’t be necessary when NHibernate manages the version on the “client side”, like:

   1: UPDATE MyEntity SET Version = 2 WHERE ...;

   2: //or

   3: UPDATE MyEntity SET Version = GETDATE() WHERE ...;

But not if the version is handled on the database side, like when using SQL Server’s ROWVERSION/TIMESTAMP columns or Oracle’s ORA_ROWSCN pseudo-columns. In these cases, NHibernate needs to issue a SELECT after each INSERT or UPDATE:

   1: UPDATE MyEntity SET ... WHERE ...;

   2: SELECT Version FROM MyEntity WHERE ...;

This breaks batching, because it needs to be done immediately after each INSERT/UPDATE.