Entity Framework Contains() only if non-null int

I think this is right but want to check as I'm still not totally comfortable with EF and so am still sometimes writing non-optimal queries.

For the classes:

public class Country : IOrganization
{
	public int Id { get; private set; }

	public ICollection<Campaign>? Campaigns { get; private set; }
}

public class Campaign : IOrganization
{
	public int Id { get; private set; }

	public State? State { get; private set; }
	public int? StateId { get; set; }

	public Country? Country { get; private set; }
	public int? CountryId { get; set; }
}

Every campaign is owned by a State or a Country. Never both, never neither.

For a list of Countries (that I'm about to delete) I need to get the associated list of child Campaigns. So I'm doing the following, checking for a non-null CountryId and then using Contains().

List<Country> deletedCountries = ChangeTracker.Entries<Country>()
	.Where(e => e.State == EntityState.Deleted)
	.Select(e => e.Entity).ToList();

List<Campaign> campaignsIndirectDelete = await Campaigns
		.Where(campaign => campaign.CountryId != null)
		.Where(campaign => deletedCountries
			.Select(country => country.Id)
			.Contains(campaign.CountryId!.Value))
		.ToListAsync();

Is this the most efficient way to accomplish this?

Update:

The above code is in my DbContext.SaveChanges()/SaveChangesAsync(). It's there to handle deletes where Entity Framework can't correctly handle cascading deletes. (Please avoid discussion in this question about the best way to structure models so Entity Framework can cascade deletes as that is off-topic to this question.)

I need these lists to first delete a child collection in each. Then I delete the collections objects too. So that constrains what I can do.

Your approach seems to be correct, but it's overcomplicating the problem. Since you have not shared the code of the actual delete, but only the part which determines what needs to be deleted, it's difficult to comment on your actual deletion from the perspective of efficiency.

Yet, the absolute most efficient removal is that which translates to something like

delete from campaign where CountryId in (1, 2, 3)

or

update campaign set IsDeleted = 1 where CountryId in (1, 2, 3)

Yet, sometimes you do not win that much by improving efficiency. So, before you are to optimize efficiency make sure that you actually have an efficiency problem to fix.

Now, if you want to simplify your code, then you could specify

ON DELETE CASCADE

for your FOREIGN KEY constraints, so, if you are deleting the record for real (rather than soft-deleting it by marking a deleted field as true), then the RDBMS will perform the deletion automatically in the background and you will have nothing to resolve, unless you really have a performance issue.

If you have an IsDeleted field for countries, then you can also update/delete the campaigns to manually cascade your deletion. You can also use a trigger upon the change of an IsDeleted field of Country records to update the campaigns associated to it accordingly.