问题
Basically I am mixing EF with one call to a stored procedure which does some batch deletions, otherwise EF is too slow.
Here is some pseudo-code of this scenario (I have more complex code in reality):
public void RemoveCustomer(int customerUID)
{
// this code is running in one db transaction
{
// retrieve certain orders of particular customer using EF
var orders = repoOrders.GetOrdersOfCustomer(filter, customerUID);
// do something with above orders before deletion using EF
repoX.DoSomethingWithOrders(orders);
// call SP to delete all orders of customer
repoOrders.DeleteAllOrdersOfCustomer(customerUID); // this calls a stored procedure
// delete customer using EF
repoCustomers.DeleteCustomer(customerUID); // throws exception due to relationship!
}
}
of course, customers-orders is one-to-many (1:m) relation.
I want to avoid in the above scenario the exception which is thrown when there will be some orders loaded by the context belonging to the customer that gets deleted. Exception is:
"The relationship could not be changed because one or more of the foreign-key properties is non-nullable. When a change is made to a relationship, the related foreign-key property is set to a null value. If the foreign-key does not support null values, a new relationship must be defined, the foreign-key property must be assigned another non-null value, or the unrelated object must be deleted."
So, I want to know if is possible to clear some/all orders from <DbSet>.Local without causing any change into the database after calling the stored procedure and before user gets deleted.
I guess Detach could be used, but this means I should loop through.
What would you recommend?
Edit: I am new to EF and I am integrating EF now after the repositories were done using ADO.NET, and yes, the BL has been kept the same... so I try this integration with minimal efforts at this point.
Note: I can't make changes on the database structure.
回答1:
I'd personally make each repository method use its own context, save its own changes at the end of the method, etc., and then use a TransactionScope to ensure that the operations are atomic.
void DeleteAllOrdersOfCustomer(Guid customerUID)
{
using (var context = new Context())
{
...
context.SaveChanges();
}
}
...
using (var ts = new TransactionScope())
{
// call SP to delete all orders of customer
repoOrders.DeleteAllOrdersOfCustomer(customerUID); // this calls a stored procedure
// delete customer using EF
repoCustomers.DeleteCustomer(customerUID); // throws exception due to relationship!
ts.Complete();
}
回答2:
I have been using SqlQuery from dbset for this purpose, this way:
context.xdbset.SqlQuery("storedprocedure; select * from xdbset");
The query calls the stored procedure and then calls select *... which necessary here because the query result have to contain the data format expected for the context to build "xdbset" objects.
You can look here for a similar debate.
回答3:
Refresh should do the trick
It can be a little tricky and I don't know how expensive it is to use. You will need to use the RefreshMode Enum and I assume you want StoreWins.
BTW you might be able to speed up your EF delete, instead of the batch, but that is for a different question.
来源:https://stackoverflow.com/questions/24085175/ef6-is-there-a-way-to-clear-ef-cache-local-items-without-causing-any-change