I'm confused as to the difference. Being fairly new to .Net, I know I can query IEnumerables
using the Linq extensions. So what is this IQueryable
and how does it differ?
views:
822answers:
5IQueriable is the same as IEnumerable but it also provides additional functionality to implement custom querying with Linq. Here is description on MSDN: http://msdn.microsoft.com/en-us/library/system.linq.iqueryable.aspx
"The primary difference is that the extension methods defined for IQueryable take Expression objects instead of Func objects, meaning the delegate it receives is an expression tree instead of a method to invoke. IEnumerable is great for working with in-memory collections, but IQueryable allows for a remote data source, like a database or web service"
Source: here
IEnumerable<T>
represents a forward-only cursor of T
. .NET 3.5 added extension methods that included the LINQ standard query operators
like Where
and First
, with any operators that require predicates or anonymous functions taking Func<T>
.
IQueryable<T>
implements the same LINQ standard query operators, but accepts Expression<Func<T>>
for predicates and anonymous functions. Expression<T>
is a compiled expression tree, a broken-up version of the method ("half-compiled" if you will) that can be parsed by the queryable's provider and used accordingly.
For example:
IEnumerable<Person> people = GetEnumerablePeople();
Person person = people.Where(x => x.Age > 18).FirstOrDefault();
IQueryable<Person> people = GetQueryablePeople();
Person person = people.Where(x => x.Age > 18).FirstOrDefault();
In the first block, x => x.Age > 18
is an anonymous method (Func<Person, bool>
), which can be executed like any other method. Enumerable.Where
will execute the method once for each person, yield
ing values for which the method returned true
.
In the second block, x => x.Age > 18
is an expression tree (Expression<Func<Person, bool>>
), which can be thought of as "is the 'Age' property > 18".
This allows things like LINQ2SQL to exist because they are can parse the expression tree and convert it into equivalent SQL. And because the provider doesn't need to execute until the IQueryable
is enumerated (it implements IEnumerable<T>
, after all), it can combine multiple query operators (in the above example Where
and FirstOrDefault
) to make smarter choices on how to execute the entire query against the underlying data source (like using SELECT TOP 1
in SQL).
See:
In real life, if you are using a ORM like Linq-to-Sql
- If you create an IQueriable, then the query may be converted to sql and run on the database server
- If you create an IEnumerable, then all rows will be pulled into memory as objects before running the query.
The principle difference is that IEnumerable will enumerate all of its elements all the time, while IEqueryable will enumerate elements, or even do other things, based on a query. The query is an Expression (a data representation of .Net code), which an IQueryProvider must explore/interpret/compile/whatever in order to generate results.
Having a query expression gives two advantages.
The first advantage is optimization. Because modifiers like 'Where' are included in the query expression, the IQueryProvider can apply otherwise impossible optimizations. Instead of returning all elements then throwing away most of them due to a 'Where' clause, the provider could use a hash table to locate items with a given key.
The second advantage is flexibility. Because Expressions are explorable data structures, you can do things like serialize the query and send it to a remote machine (eg. linq-to-sql).