Question

This code

int maxTrackId = dataContext.TicketTracks.Max(T => T.TrackId);

Generates the following SQL code:

SELECT MAX([t0].[TrackId]) AS [value]
FROM [dbo].[TicketTracks] AS [t0]

But if the table is empty, I got an exception saing that nullable value cannot be assigned to non-nullable variable.

Then I write this code, explicitly telling the SQL provider that i need either maximum or null:

int? maxTrackId = dataContext.TicketTracks.Max(T => (int?)T.TrackId);

And it works fine on that non-nullable SQL column, giving me back either null or the number. But the SQL code generated is kind of strange:

SELECT MAX([t1].[value]) AS [value]
FROM (
    SELECT [t0].[TrackId] AS [value]
    FROM [dbo].[TicketTracks] AS [t0]
    ) AS [t1]

So it seems weird, especially with two SELECTs. Is there any way to circumvent this?

Answer 1

The SQL might be convoluted but there is no performance difference. The SQL Server query optimizer reliably optimizes both statements to the same plan.

Really, removing a redundant nested select is a trivial case. I have never seen this cause a plan difference.

So it is an aesthetic problem, not a practical one. ORMs generate convoluted code and I learned to live with it.