LINQ to NHibernate does not JOIN with subclass, whereas QueryOver does - nhibernate

I had this QueryOver query:
var result = session.QueryOver<TopicSelection>()
.Where(x => x.LacId == lac && x.RemoveDate == null)
.List();
In a nutshell: TopicSelection is a base class, and one subclass has many-to-one property, with lazy=false and fetch=join.
When I used QueryOver, NHibernate created nice join and fetched additional data from many-to-one table. Which is great, one query is issued and I get everything.
When I changed it to LINQ to NHibernate:
var result = session.Query<TopicSelection>()
.Where(x => x.LacId == lac && x.RemoveDate == null)
.ToList();
executed query does not contain JOIN. What is more, every time a many-to-one property is needed, an additional select is issued.
Is is a bug in LINQ in NHibernate? Can I instruct LINQ to NHibernate query to fetch data for subclass?

What you are experiencing is a "NHibernate LINQ implementation" as is today. We can effectively do two things.
In case, that we can adjust the query to ask for a subclass, or the many-to-one property is (could be) declared on base TopicSelection, we can use .Fetch()
var result = session
//.Query<TopicSelection>()
.Query<TopicSelectionSubClass>()
.Where(x => x.LacId == lac && x.RemoveDate == null)
.Fetch(x => x.Category ) // the many-to-one property joined
.ToList();
The second approach is even better. I would suggest that almost in any case. I.e. instead of any many-to-one mapping with lazy="false" fetch="join" - let's use:
19.1.5. Using batch fetching
Change the class mapping like this:
<class name="Category" batch-size="25" ...
That will change the 1+N select into 1+1 (or 1+2). Small cite from docs:
NHibernate can make efficient use of batch fetching, that is, NHibernate can load several uninitialized proxies if one proxy is accessed (or collections). Batch fetching is an optimization of the lazy select fetching strategy. There are two ways you can tune batch fetching: on the class and the collection level.
Batch fetching for classes/entities is easier to understand. Imagine you have the following situation at runtime: You have 25 Cat instances loaded in an ISession, each Cat has a reference to its Owner, a Person. The Person class is mapped with a proxy, lazy="true". If you now iterate through all cats and call cat.Owner on each, NHibernate will by default execute 25 SELECT statements, to retrieve the proxied owners.
Read more here

Related

NHibernate Queryover - Is there any way of getting better SQL out of nHibernate when retrieving this object plus collections graph?

We have a situation where we are trying to retrieve a couple of levels deep for an object graph using QueryOver. So if our top level is ParentEntity and our children are ChildEntitysA, ChildEntitysB and ChildEntitysC then we have something like the following code to retrieve our graph.
var query = session.QueryOver<ParentEntity>(() => pAlias).Where(() => pAlias.Id == key).Future<ParentEntity>();
var queryA = session.QueryOver<ParentEntity>(() => pAlias).Left.JoinAlias(x => x.ChildEntitysA, () => caAlias).Where(() => pAlias.Id == key).Future<ParentEntity>();
var queryB = session.QueryOver<ParentEntity>(() => pAlias).Left.JoinAlias(x => x.ChildEntitysB, () => cbAlias).Where(() => pAlias.Id == key).Future<ParentEntity>();
var queryC = session.QueryOver<ParentEntity>(() => pAlias).Left.JoinAlias(x => x.ChildEntitysC, () => ccAlias).Where(() => pAlias.Id == key).Future<ParentEntity>();
return query.ToList();
This should generate 4 SQL statements in one single call to the database and brings back what we want. However, the SQL it generates does have some inefficiency as all of the child queries will contain in the SELECT statement all the columns of the parent as well as the child entity's columns, something akin to:
SELECT parent.col0, parent.col1, parent.col2, parent.col3, .... parent.col99
FROM parent WHERE .....
SELECT parent.col0, parent.col1, parent.col2, parent.col3, .... parent.col99, childA.col0, childA.col1, childA.col2 .... childA.col99
FROM parent LEFT OUTER JOIN childA ON ....
SELECT parent.col0, parent.col1, parent.col2, parent.col3, .... parent.col99, childB.col0, childB.col1, childB.col2 .... childB.col99
FROM parent LEFT OUTER JOIN childB ON ....
SELECT parent.col0, parent.col1, parent.col2, parent.col3, .... parent.col99, childC.col0, childC.col1, childC.col2 .... childC.col99
FROM parent LEFT OUTER JOIN childC ON ....
yet the query generated for the parent entity already contains all the necessary data! This does become a bit of an efficiency concern if we're returning multiples of the parent and the volume of data per child query rises accordingly (especially as the entities involved are quite large in terms of the number of underlying columns).
So, is there any way we can force nHibernate to not generate SQL to return all the parent entity's values for every single part of the query and just limit it to returning minimal key columns only in the SQL? Is Queryover even the optimal API for this situation?
The way I would suggest, is to use full power of NHibernate features in these combinations:
The first is a complex query.
The second is to profit from lazy loading and batch-size setting
So, in the first case we are about to create one and only one QueryOver (maybe including some subqueries)
It must contain projections, that's essential. it can include (many) Left, Inner joins, can be deeply filtered... We will then get the narrowed SELECT statement, containing only required fields. This results in one SQL query, which must be converted into some (unmapped) DTO object
The Projections and DTO means, that at the end we have to use ResultTransformer, which will convert all coming selected fields into DTO object properties.
In the second case, we are trying to profit from lazy behaviour and batch-size mapping. It could look like this:
1) class/ entity
<class name="ParentEntity" batch-size="25" lazy="true">
...
2) collection
<bag name="Children" lazy="true" batch-size="25"
...
See more in documentation: 19.1.5. Using batch fetching
What we gain here is, that we can create soft queries returning only light objects (these which were mapped). This is the first Select. Then all the properties (many-to-one, one-to-many) are loaded in batches - but only if they are accessed, needed.
This leads to more than 1 SELECT clause, but also there is no inflation of SELECT clauses like 1 + N. If NHibernate will find out, that some of the required properties is already loaded in the session... it won't fire the SELECT any more.
Summary: both approaches are really the way. Try to play arround to find out in which situation you get more from the first or the second. BUT the mapping with lazy="true" and batch-size="25" should be used anyway
Some more links about batch-size:
How to Eager Load Associations without duplication in NHibernate?
NHibernate QueryOver with Fetch resulting multiple sql queries and db hits
Is this the right way to eager load child collections in NHibernate

Is there a way to override a collection's mapped order-by clause when paging in NHibernate?

I have an entity with a bag collection and I want to page through the entire data set. The collection has an order-by clause. When I try to eager-load the collection, NHibernate is generating SQL that is very very slow because it causes SQL Server to sort on a very non-unique property, and there is a large amount of data in the table.
The code:
var session = NHibernateSessionManager.Instance.GetSession();
session.CreateCriteria<Album>()
.SetFetchMode("Track", FetchMode.Eager)
.SetMaxResults(1000)
.SetFirstResult(1)
.AddOrder(new Order("Id", true))
.List();
The bag mapping (note the order-by attribute):
<bag name="Track" inverse="true" lazy="true" batch-size="1000" cascade="none"
order-by="TrackNumber ASC">
The relevant SQL generated:
OVER(ORDER BY track2_.TrackNumber, this_.Id) as __hibernate_sort_row
If i remove the order-by from the mapping then the SQL changes to this (much better):
OVER(ORDER BY this_.Id) as __hibernate_sort_row
So the question is: is there a way to override or remove the mapped order-by clause?
After comments from Thilak Nathen and Firo, it looks like the short answer to the question is: "No you can't do that". The long answer is to remove the order-by attribute from the mapping and add it as needed. The reason it was there is that it's always needed, except for this one unusual situation. Firo suggested overriding the factory configuration which I think will work because this is for reporting. For anyone else who wants to know how to do it, here's the code to remove an order-by from an xml mapping programmatically:
Configuration cfg = new Configuration().Configure();
var albumMapping = cfg.ClassMappings.Where(x => x.DiscriminatorValue == "Foo.Album").First();
var trackProperty = albumMapping.GetProperty("Track");
var bagMapping = ((NHibernate.Mapping.Bag)trackProperty.Value);
bagMapping.OrderBy = "";
sessionFactory = cfg.BuildSessionFactory();

Unable to figure out how to do Joins within IQueryable

Here is what I am trying:
IQueryable query = this.MyRepository.GetShippingCollection();
IList<SomeListType> myList = query.Where(x => x.Settings
.Where(y => y.SelectorID.Equals(5))
.Count() > 0)
.OrderBy(x => x.Order)
.ToList();
Produces this error:
could not resolve property: Settings.ID
If I do it this way it works, but causes over 3,000 queries on my SQL Server:
IList<SomeListType> myList = this.MyRepository.GetShippingCollection().ToList();
myList = myList.Where(x => x.Settings
.Where(y => y.SelectorID.Equals(5))
.Count() > 0)
.OrderBy(x => x.Order)
.ToList();
I know the solution resides within using a "Join".
I have been looking at examples for the last couple hours and can only find Join examples within the Mapping file. I am also finding examples for "ICriteria".
I don't want to have to create seporate entries for all my complex queries in the mapping file so the join within that file will not work.
Since I am using Fluent NHibernate, I am not using "ICriteria". I am using "IQueryable". How do you join multiple tables within "IQueryable"?
Any help with this would be greatly appreciated.
Thanks in advance.
If the second query is executing 3,000 queries, it is almost certainly lazy-loading the Settings collection. As you iterate over the list, you access this collection, and each time NHibernate goes back to the database to fetch it. Try setting the fetch mode for the Settings property to eager load in the mapping.
Beyond that, the LINQ provider could be an issue. What version of NHibernate are you using? The 2.x LINQ provider has some real limitations. It has been reimplemented in the 3.0 trunk, but you'll have to download and compile it from the source.
By the way, ICriteria vs IQueryable is not related to Fluent NHibernate. Criteria API and LINQ are two providers through which you can create queries. Fluent NHibernate is an alternative way to perform configuration.

How do I use Fluent Nhibernate many-to-many for optimal performance?

I have a product table that has a many-to-many relation to itself (using a two-column many-to-many table) and I have set it up in Fluent NHibernate with the following code:
public class ProductConfiguration : ClassMap<Product>
{
public ProductConfiguration()
{
Table("Product");
Id(p => p.Id).GeneratedBy.Guid();
Map(p => p.Name).Not.Nullable().Length(254);
Map(p => p.Description).Not.Nullable().Length(1000);
Map(p => p.CreatedAt).Not.Nullable();
HasManyToMany(p => p.CrossSell)
.Table("ProductCrossSell")
.ParentKeyColumn("Id")
.ChildKeyColumn("ProductId");
}
}
My MVC application has two pages that uses this setup:
Index - Uses a generic repository GetAll method to display all products.
Detail - Uses a generic repository GetById method to display one product and any related cross sell products setup in the many-to-many realation.
It looks like NHibernate is set to LazyLoad the many-to-many by default so when I fire up the application and watch it in profiler I can see that it does LazyLoad the many-to-many with the following alert "Use of implicit transactions is discouraged".
How do I get rid of this alert? I couldn't find any information on how to wrap a LazyLoad inside a transaction to get rid the alert. Is it even possible?
Is there a way to not lazyload this by telling NHibernate that whenever I ask for GetById make sure to join the tables a get everything in one query? I tried using .Fetch.Join() in the many-to-many mapping but that also affected my GetAll query which now displays a joined result set as well which is incorrect.
What is the best apprach for this kind of simple scenario?
Thanks
The way to get rid of the warning is to access the object graph and fully populate the UI elements inside a single transaction.
Not by configuration. You can create an HQL query that eager fetches the association and use that query for a specific view. I would stick with lazy loading and not make that optimization unless needed. The HQL would be:
return session.CreateQuery("from ProductionConfiguration pc join fetch pc.CrossSell where pc.Id = ?")
.SetGuid(0, id)
.List<ProductConfiguration>();
All collections are lazily loaded in NHibernate by default.
You must be triggering loading with a call of some kind (maybe even with the debugger watches)

NHibernate - Incorrect thinking? Subclassed Model based on Join

I have a simple model class (Part), which pulls from it's information from a single table (t_Part).
I would like a subclass of this model called (ProducedPart), that would still utilize NHibernate's caching mechanisms, but would only be instances of (Part) that have a foreign key relationship in a table called "t_PartProduction". I do not need to have a model for this second table.
I only need a read-only version of ProducedPart
I could always implement a Facade/Repository over this, but I was hoping to setup a mapping that would pull "t_Part" joined with "PartProduction" when I asked for "ProducedPart" in NH.
Is this the wrong way to use NH?
Edit
So, the SQL would look something like
SELECT p.*
FROM t_Part p
INNER JOIN t_PartProduction pp ON pp.PartID = p.PartID
WHERE pp.ProductionYear = '2009'
I believe what you are looking for is a joined subclass. In FNH, it will look something like:
public class PartMap : ClassMap<Part>
{
public PartMap()
{
Id(x => x.Id)
JoinedSubClass<ProducedPart>("PartID", sub => {
sub.Map(x => x.Name);
sub.Map(x => x.ProductionYear);
});
}
}
In order have NHibernate cache the results, you will need to have the subclass mapped (and if you didn't map it, you wouldn't be able to get NH to load it in the first place).
Bringing in some context from the FNH groups thread, it will not explicitly be read-only though. In my opinion, making things read-only is not an appropriate thing for NHibernate to manage. This is better controlled by the database and connections (i.e. creating a connection to the database that only has SELECT permissions on the tables/views being accessed). See my answer to a previous SO question about readonly sessions in NHibernate for more of my thoughts on the matter.
The key here is using both the where and mutable elements of the class definition for NHibernate Mappings.
Using Fluent NHibernate, this looks like:
public Part()
{
WithTable("t_Part");
Id(i => i.Id).ColumnName("PartID");
Map(m => m.Name).ColumnName("Part");
SetAttribute("where", "PartID IN ( SELECT pp.PartID FROM t_PartProduction pp WHERE pp.ProductionYear = '2009' ) ");
ReadOnly();
}
No, this is perfectly possible. Look in the NHibernate documentation for the "table per subclass" model of inheritance. It will actually implement this as a LEFT JOIN, so that when you load a Part, it creates an instance of either your Part or your ProducedPart class depending on whether the other row is present. You'll find documentation on nhibernate.info.
I'm not sure you could make ProducedPart read-only doing this though.
I'm assuming from this:
WHERE pp.ProductionYear = '2009'
that you want the subclass only where the production year is 2009, i.e. if when there is a record in t_PartProduction for a different year, you want this Part treated as a plain Part object, not a ProducedPart, then you could consider creating a view definition within your database that is a filtered version of t_PartProduction, then making your subclass join to this view rather than the base table.