If I have a model that looks like this:
and I do a Linq to Entities query like this:
var c = MyContext.Contact.Count();
I'll get a SQL query that is as big as all out doors!
SELECT
[GroupBy1].[A1] AS [C1]
FROM ( SELECT
COUNT(1) AS [A1]
FROM [dbo].[Contacts] AS [Extent1]
LEFT OUTER JOIN (SELECT
[UnionAll1].[Id] AS [C1]
FROM (SELECT
[Extent2].[Id] AS [Id]
FROM [dbo].[Companies] AS [Extent2]
UNION ALL
SELECT
[Extent3].[Id] AS [Id]
FROM [dbo].[Employees] AS [Extent3]) AS [UnionAll1]
UNION ALL
SELECT
[Extent4].[Id] AS [Id]
FROM [dbo].[Contractors] AS [Extent4]) AS [UnionAll2] ON [Extent1].[Id] = [UnionAll2].[C1]
LEFT OUTER JOIN (SELECT
[UnionAll3].[Id] AS [C1]
FROM (SELECT
[Extent5].[Id] AS [Id]
FROM [dbo].[Suppliers] AS [Extent5]
UNION ALL
SELECT
[Extent6].[Id] AS [Id]
FROM [dbo].[Customers] AS [Extent6]) AS [UnionAll3]
UNION ALL
SELECT
[Extent7].[Id] AS [Id]
FROM [dbo].[People] AS [Extent7]) AS [UnionAll4] ON [Extent1].[Id] = [UnionAll4].[C1]
) AS [GroupBy1]
It seems to me that this should be a very simple query that executes over the base type table (Contact in this case) The model that I've included here is a watered down sample of what I'm working with. As you can imagine with a hierarchy 6 levels deep rather than 3 the SQL queries on anything other than the most derived types are very expensive.
Is there any way to tweek the query, or the model definition to reduce the unnecessary complexity of this query.
See Question&Answers more detail:
os 与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…