Question of LINQ optimisation

If I have the following code, is the compiler instanti开发者_运维知识库ating each result or is it wise enough to just count how many corresponding records are in the table? If not, it may force me to use a different strategy on larger queries.

from c in context.RendezVous
where c.RepID == repID &&
    c.DateHeureRV != null &&
    c.DateHeureRV.Value.Date == date.Date
select c).Count();

Thank you!

---

It depends on the type of context.

If this is an Entity Framework or Linq to SQL query, and context is IQueryable<T>, then the query gets turned into a SQL query on the server which just returns the count as a single integer.

If this is an in-memory collection (ie: IEnumerable<T>), each item is iterated in sequence (Linq to Objects) and counted.

I suspect the former is true, since you mentioned "table" and you're not using the LINQ to Dataset extension methods. In that case, you will stay very efficient.

---

This will execute (more or less):

select count(*)
from rendezvous
where repid='...' and dateheurerv is not null and dateheurerv='...'

Linq will add no real overhead to the query and the count will be processed server-side.

---

This will execute a SQL query with the three WHERE conditions you listed and return the count of the matching rows, it will only materialize the actual count, not each row - if that was your question.