I have some Entity Framework code that I don't think is very efficient. Basically, I have an ApplicationUser class which looks like this:
public class ApplicationUser : IdentityUser
{
public string CompanyName { get; set; }
public virtual ICollection<SolverRun> Runs { get; set; }
// .. other stuff generated by the Visual Studio template
}
What I want to do is create a new row in SolverRuns
for the current user:
// Log run in database
var manager = new UserManager<ApplicationUser>(new UserStore<ApplicationUser>(new ApplicationDbContext()));
var currentUser = manager.FindById(OwnerId);
currentUser.Runs.Add(new SolverRun
{
Base = job.Base,
Duration = context.JobRunTime,
Result = (context.Result ?? "").ToString(),
RunName = job.RunName,
Start = context.FireTimeUtc.GetValueOrDefault().LocalDateTime
});
manager.Update(currentUser);
This works great. However, what I'm seeing is the following query:
exec sp_executesql N'SELECT
[Extent1].[SolverRunID] AS [SolverRunID],
[Extent1].[RunName] AS [RunName],
[Extent1].[Base] AS [Base],
[Extent1].[Start] AS [Start],
[Extent1].[Duration] AS [Duration],
[Extent1].[Result] AS [Result],
[Extent1].[ApplicationUser_Id] AS [ApplicationUser_Id]
FROM [dbo].[SolverRuns] AS [Extent1]
WHERE ([Extent1].[ApplicationUser_Id] IS NOT NULL) AND ([Extent1].[ApplicationUser_Id] = @EntityKeyValue1)',N'@EntityKeyValue1 nvarchar(128)',@EntityKeyValue1=N'029a20a9-b487-4579-87d0-50608ed7f058'
Which will select every row in the SolverRuns
table for that user. I suspect that's happening on the getter for currentUser.Runs
.
I'm concerned that as a user adds more and more runs (there could eventually be thousands), this will get slower and slower. Is there a way to just add the new row, and not select everything first? Thanks!
Update:
Based on py3r3str's answer below, here's my new working code (after changing the model and migrating the database and what not):
// Log run in database
using (var dbContext = new ApplicationDbContext())
{
var manager = new UserManager<ApplicationUser>(new UserStore<ApplicationUser>(dbContext));
var currentUser = manager.FindById(job.OwnerId);
dbContext.Set<SolverRun>().Add(new SolverRun
{
Base = job.Base,
Duration = context.JobRunTime,
Result = (context.Result ?? "").ToString(),
RunName = job.RunName,
Start = context.FireTimeUtc.GetValueOrDefault().LocalDateTime,
ApplicationUser = currentUser
});
dbContext.SaveChanges();
}
It now will run this INSERT
statement (which is odd, but I gave up trying to understand EF SQL queries so I'll assume someone smarter than me thinks it's efficient):
exec sp_executesql N'INSERT [dbo].[SolverRuns]([ApplicationUserId], [RunName], [Base], [Start], [Duration], [Result])
VALUES (@0, @1, @2, @3, @4, @5)
SELECT [SolverRunID]
FROM [dbo].[SolverRuns]
WHERE @@ROWCOUNT > 0 AND [SolverRunID] = scope_identity()',N'@0 nvarchar(128),@1 nvarchar(max) ,@2 nvarchar(max) ,@3 datetime2(7),@4 time(7),@5 nvarchar(max) ',@0=N'029a20a9-b487-4579-87d0-50608ed7f058',@1=N'PilotRun',@2=N'YUL',@3='2015-09-03 11:07:32.0353181',@4='00:00:04.1521401',@5=N'Success'
Note that calling currentUser.Runs
will lazily load all runs, as expected.