OData $orderby clause on collection property - asp.net-web-api

I have the following classes :
public class Parent
{
public string ParentProp { get; set; }
public IEnumerable<Child> ManyChildren { get; set; }
}
public class Child
{
public string ChildName { get; set; }
public int Value { get; set; }
}
Say I have an OData operation defined which returns IEnumberable<Parent>. Can I write an $orderby clause which performs the following operation ('parents' is an IEnumerable<Parent>) :
parents.OrderBy(x => x.ManyChildren.Single(y => y.ChildName == "Child1").Value);
I know I can write custom actions (http://msdn.microsoft.com/en-us/library/hh859851(v=vs.103).aspx) to do this ordering for me, but I'd rather use an $orderby clause.
(The only SO question which asked something similar is a little dated - How can I order objects according to some attribute of the child in OData?)

As I tried is possible with nesting $orderby in $expand so will be:
odata/User?&$select=Active,Description,Name,UserId&$expand=Company($select=Active,Name,CreatedBy,CompanyId;$orderby=Active asc)
And what you get is somthing like:
ORDER BY [Project2].[UserId] ASC, [Project2].[C19] ASC
will order a company collection for each user separately.
I think in version OData Client for .NET 6.7.0 is supported, in release notes is writhing:
In query options
$id, $select, $expand(including nested query options)....
I see in version 6.1 that values for nested options exist and is in:
DataQueryOptions->SelectExpand->SelectExpandClasue->SelectedItems->ExpandNavigationItem->OrderByOption
but is not working.
I tried and with System.Web.OData 5.6 and all releated dependencies but seams is not working.
My conclusion:
Seams that is everiting prepared like DataQueryOptions exist nested orderby but is not working.
Like I find out standard seams is going in that direction.
https://issues.oasis-open.org/browse/ODATA-32

It depends on your OData service implementation. Which kind of service are you using? WCFDS, WebAPI, or the service you implement yourself?
Url parser do can parse the URL such as root/People?$orderby=Company/Name. The translator is implemented by service.
And I agree with the answer in related question: "it's not possible to do this with a navigation property that has a cardinality of many". Since it's has a cardinality of many, service cannot know which one should be used to sorting.

Related

OData EDM model - auto-expanding a nested Entity Type

I'm using OData v4 and the models are configured with the EdmModel Builder for our .NET Web API.
I have two models defined like this:
public class Customer
{
public int CustomerId { get; set; }
public string CustomerName { get; set; }
}
public class Order
{
public int OrderId { get; set; }
public string Category { get; set; }
public Customer OrderCustomer { get; set;}
}
These models have corresponding controllers and are registered as follows:
builder.EntitySet<Customer>("Customers")
.EntityType
.HasKey(x => x.CustomerId);
builder.EntitySet<Order>("Orders")
.EntityType
.HasKey(x => x.OrderId)
.Expand(
maxDepth: 2,
expandType: SelectExpandType.Automatic,
properties: nameof(Order.OrderCustomer));
I'm able to make OData requests to both these endpoints as follows:
/Customers/{Id} and
/Orders/{Id}.
I'm expecting that when I query the Orders, the nested EntitySet Customers will auto expand since I've set expandType: SelectExpandType.Automatic on the Order EntitySet. However, I can't get the CustomerOrder property to auto-expand on the Customer and I have to call the request with an expand parameter:
/Orders/{Id}?$expand=OrderCustomer.
I think this is because both Customer and Order are registered as EntitySets, so OData expects that an expand parameter be provided if they are nested. Is there a way to get the OrderCustomer property to auto-expand (i.e. without the need for the expand parameter to be provided)? My understanding of OData/ Edm Models is pretty elementary so any help is appreciated.
Your fluent configuration is correct, for OData v4, that will work for both collection and item queries.
If it is not working for you there are 3 possible issues:
You do not appear to be using the OData v4 URL convention for item queries, in v4 the expected URL is:
/Orders({Id})
This brings into question how you modified the router to support the v3 syntax, there are multiple variation on how to implement the v3 routes so it is possible that changes made in this area could affect the way that default expansion and selection is applied, or if it should apply.
You may not be including the navigation data in your data query. If the data is not retrieved from the data store, then it stands to reason that it will not be in the output recordset. If you are manually using the ODataQueryOptions.ApplyTo() to apply the user request to your query, then this will not take into account the configuration on the model, it will only apply thequery options that the caller has specified.
The caller might be specifying an Empty $expand= which will cancel out the auto configuration. Even if the originating caller has not specified any query options it is common enough for OData APIs to have standard or custom middleware running that might be manipulating the request query strings. To verify the URL is untampered, log it in your GET method handler and make sure the $expand is not specified.
As with the previous point, the ODataQueryOptions parameter in your GET method should NOT show any value for the SelectExpand if you want the auto configuration to be applied.
Finally, the last place to check is that you haven't overriden the default EnableQueryAttribute. If you have implemented your own custom implementation of EnableQueryAttribute then make sure that you still call the base implementation to correctly apply the ODataQueryOptions AND the schema defaults to the underlying IQueryable result.
In addition to the answer provided by Chris Schaller.
Another issue I had was caused by the casing on the property. For example, I have camelCasing enabled on the builder
builder.EnableLowerCamelCase();
This means the naming in the Expand on the configuration needed to be updated to match.
.Expand(
maxDepth: 2,
expandType: SelectExpandType.Automatic,
properties: "orderCustomer"); // <-- camelCase
This seems to be required even if EnablePropertyNameCaseInsensitive is enabled in the ODataOptions.

How to use a Dictionary or Hashtable for LINQ query performance underneath an OData service

I am very new to OData (only started on it yesterday) so please excuse me if this question is too dumb :-)
I have built a test project as a Proof of Concept for migrating our current web services to OData. For this test project, I am using Reflection Providers to expose POCO classes via OData. These POCO classes come from in-memory cache. Below is the code so far:
public class DataSource
{
public IQueryable<Category> CategoryList
{
get
{
List<Category> categoryList = GetCategoryListFromCache();
return categoryList.AsQueryable();
}
}
// below method is only required to allow navigation
// from Category to Product via OData urls
// eg: OData.svc/CategoryList(1)/ProductList(2) and so on
public IQueryable<Category> ProductList
{
get
{
return null;
}
}
}
[DataServiceKeyAttribute("CategoryId")]
public class Category
{
public int CategoryId { get; set; }
public string CategoryName { get; set; }
public List<Product> ProductList { get; set; }
}
[DataServiceKeyAttribute("ProductId")]
public class Product
{
public int ProductId { get; set; }
public string ProductName { get; set; }
}
To the best of my knowledge, OData is going to use LINQ behind the scenes to query these in-memory objects, ie: List in this case if somebody navigates to OData.svc/CategoryList(1)/ProductList(2) and so on.
Here is the problem though: In the real world scenario, I am looking at over 18 million records inside the cache representing over 24 different entities.
The current production web services make very good use of .NET Dictionary and Hashtable collections to ensure very fast look ups and to avoid a lot of looping. So to get to a Product having ProductID 2 under Category having CategoryID 1, the current web services just do 2 look ups, ie: first one to locate the Category and the second one to locate the Product inside the Category. Something like a btree.
I wanted to know how could I follow a similar architecture with OData where I could tell OData and LINQ to use Dictionary or Hashtables for locating records rather than looping over a Generic List?
Is it possible using Reflection Providers or I am left with no other choice but to write my custom provider for OData?
Thanks in advance.
You will need to process expression trees, so you will need at least partial IQueryable implementation over the underlying LINQ to Objects. For this you don't need a full blown custom provider though, just return you IQueryable from the propties on the context class.
In that IQueryable you would have to recognize filters on the "key" properties (.Where(p => p.ProductID = 2)) and translate that into a dictionary/hashtable lookup. Then you can use LINQ to objects to process the rest of the query.
But if the client issues a query with filter which doesn't touch the key property, it will end up doing a full scan. Although, your custom IQueryable could detect that and fail such query if you choose so.

Linq dynamic queries for user search screens

I have a database that has a user search screen that is "dynamic" in that I can add additional search criteria on the fly based on what columns are available in the particular view the search is based on and it will allow the user to use them immediately. Previously I had been using nettiers for this database, but now I am programming a new application against it using RIA and EntFramework 4 and LINQ.
I currently have 2 tables that are used for this, one that fills the combobox with the available search string patterns:
LastName
LastName, FirstName
Phone
etc....
then I have an other table that splits those criteria out and is used in my nettiers algorithms. It works well, but I want to use LINQ..and it doesnt fit this model very well. Besides I think I can pare it down to just one table with linq...
using a format similar to this or something very close...
ID Criteria WhereClause
1 LastName 'Lastname Like '%{0}%'
now I know this wont fit specifically into a linq query..but I am trying to use a univeral syntax for clarity here...
the real where clause would look something like this: a=>a.LastName.Contains("{0}")
My first question is: Is that even possible to do? Feed a lambda in to a string and use it in a Linq Query?
My second question is: at one point when I was researching this before I found a linq syntax that had a prefix like it.LastName{0}
and I appear to have tried using it because vestiges of it are still in my test databases...but I dont know recall where I read about it.
Is anyone doing this? I have done some searches and found similar occurances but they mostly have static fields that are optional, not exactly the way I am doing it...
As for your first question, you can do this using Dynamic Linq as described by Scott Gu here
var query = Northwind.Products.Where("Lastname LIKE "test%");
I'm not sure how detailed your dynamic query needs to be, but when I need to do dynamic queries, I create a class to represent filter values. Then I pass that class to a search method on my repository. If the value for a field is null then the query ignores it. If it has a value it adds the appropriate filter.
public class CustomerSearchCriteria{
public string LastName { get; set; }
public string FirstName { get; set; }
public string PhoneName { get; set; }
}
public IEnumberable<Customer> Search(CustomerSearchCriteria criteria){
var q = db.Customers();
if(criteria.FirstName != null){
q = q.Where(c=>c.FirstName.Contains(criteria.FirstName));
}
if(criteria.LastName!= null){
q = q.Where(c=>c.LastName.Contains(criteria.LastName));
}
if(criteria.Phone!= null){
q = q.Where(c=>c.Phone.Contains(criteria.Phone));
}
return q.AsEnumerable();
}

Dynamic Linq Search Expression on Navigation Properties

We are building dynamic search expressions using the Dynamic Linq library. We have run into an issue with how to construct a lamba expression using the dynamic linq library for navigation properties that have a one to many relationship.
We have the following that we are using with a contains statement-
Person.Names.Select(FamilyName).FirstOrDefault()
It works but there are two problems.
It of course only selects the FirstOrDefault() name. We want it to use all the names for each person.
If there are no names for a person the Select throws an exception.
It is not that difficult with a regular query because we can do two from statements, but the lambda expression is more challenging.
Any recommendations would be appreciated.
EDIT-
Additional code information...a non dynamic linq expression would look something like this.
var results = persons.Where(p => p.Names.Select(n => n.FamilyName).FirstOrDefault().Contains("Smith")).ToList();
and the class looks like the following-
public class Person
{
public bool IsActive { get; set;}
public virtual ICollection<Name> Names {get; set;}
}
public class Name
{
public string GivenName { get; set; }
public string FamilyName { get; set; }
public virtual Person Person { get; set;}
}
We hashed it out and made it, but it was quite challenging. Below are the various methods on how we progressed to the final result. Now we just have to rethink how our SearchExpression class is built...but that is another story.
1. Equivalent Query Syntax
var results = from person in persons
from name in person.names
where name.FamilyName.Contains("Smith")
select person;
2. Equivalent Lambda Syntax
var results = persons.SelectMany(person => person.Names)
.Where(name => name.FamilyName.Contains("Smith"))
.Select(personName => personName.Person);
3. Equivalent Lambda Syntax with Dynamic Linq
var results = persons.AsQueryable().SelectMany("Names")
.Where("FamilyName.Contains(#0)", "Smith")
.Select("Person");
Notes - You will have to add a Contains method to the Dynamic Linq library.
EDIT - Alternatively use just a select...much more simple...but it require the Contains method addition as noted above.
var results = persons.AsQueryable().Where("Names.Select(FamilyName)
.Contains(#0", "Smith)
We originally tried this, but ran into the dreaded 'No applicable aggregate method Contains exists.' error. I a round about way we resolved the problem when trying to get the SelectMany working...therefore just went back to the Select method.

Does NHibernate LINQ support ToLower() in Where() clauses?

I have an entity and its mapping:
public class Test
{
public virtual int Id { get; set; }
public virtual string Name { get; set; }
public virtual string Description { get; set; }
}
public class TestMap : EntityMap<Test>
{
public TestMap()
{
Id(x => x.Id);
Map(x => x.Name);
Map(x => x.Description);
}
}
I'm trying to run a query on it (to grab it out of the database):
var keyword = "test" // this is coming in from the user
keyword = keyword.ToLower(); // convert it to all lower-case
var results = session.Linq<Test>
.Where(x => x.Name.ToLower().Contains(keyword));
results.Count(); // execute the query
However, whenever I run this query, I get the following exception:
Index was out of range. Must be non-negative and less than the size of the
collection. Parameter name: index
Am I right when I say that, currently, Linq to NHibernate does not support ToLower()? And if so, is there an alternative that allows me to search for a string in the middle of another string that Linq to NHibernate is compatible with? For example, if the user searches for kap, I need it to match Kapiolani, Makapuu, and Lapkap.
I had this happen recently. I can tell you that ToLower() does not work and that Contains() and StartsWith() do work and are not case sensitive. You can get the desired affect by using Contains() and StartsWith() directly.
There seems to be a lot of confusion around this subject.
The "old" Linq provider (for NHibernate 2.x) probably might not support this. If that's the case, it never will because it's not maintained anymore.
The new provider (included with NHibernate 3.x) does support it (although ToUpper and ToLower seem to be inverted, see http://groups.google.com/group/nhibernate-development/browse_thread/thread/a167216e466b3241)
Contains and StartsWith map to the LIKE operator in SQL. They are not case insensitive themselves; it's the collation that makes them case insensitive, so that depends on how your column/schema were created.
Update (2010-04-09): bug confirmed and patch submitted, see https://nhibernate.jira.com/browse/NH-2169
Update (2010-05-21): patch was applied on 2010-05-01 and works as expected now.
According to the comments in these two blog posts this functionality is not implemented yet.
You might want to confirm whether the database uses case sensitivity.
If it doesn't, then you don't need .ToLower()
The accepted answer mentions using Contains() and StartsWith() which are good. but wouldn't work in cases when you want to be sure both strings are the same.
Using "==" will suffice since it is also case-insensitive. So, you no longer need to use ToLower() nor ToUpper();

Resources