I'm trying to determine whether an existing collection already contains an entity before adding (to prevent dupes). I'm "merging" two entity collections, so the entities to be compared won't be exactly identical, however I can determine a dupe based on a number of attributes (ContactID and RoleID in the example below). I see the "contains" method but I'm unclear how that method determines a match. Is it on all fields of an entity, or only the primary key? The collection to be merged is the resolver of a m:n relationship:

Contact -ID -Name ...

Project -ID -Name ...

m:n -ID -ProjectID -ContactID -RoleID

If the contact is already performing a given role on a given project then I don't want to add that entity again. Is "contains" the right place to be doing this or should I be doing it some other way?

Contains is a CollectionBase method which uses 'IndexOf' and IndexOf is a CollectionBase method which performs a linear search using Equals.

So it performs entity.Equals(toFind);.

Equals on an entity performs a PK field value compare. If one or both of the entities are new, an instance compare is performed.

Otis wrote:

Equals on an entity performs a PK field value compare. If one or both of the entities are new, an instance compare is performed.

That's what I suspected - thanks for the confirmation.

So how would I go about doing a contains/find of an entity within an existing collection matching on a subset of the entities attributes?

Just use the PK check, thus IndexOf (or contains). As that's what you're interested in I think: if an entity is already there or not.

The primary keys will be different (since I'm merging two entity collections) but there will be some fields that, as defined by the business, must not be duplicated in a collection. In the example above the same contact can't be listed twice with the same role on the same project. i.e. together the projectid, contactid, and roleid must be unique. I am merging the entitycollections between projects so I know that the projectid will differ, but I want to know whether there's already an entity in the collection with the same contactid and roleid before adding an entity to the collection. I guess I could add a unique constraint on the table (which I don't currently have) and then just try adding all the entities and catching the exception, but this seems a little odd?

Still :

ChicagoKiwi wrote:

The primary keys will be different (since I'm merging two entity collections) but there will be some fields that, as defined by the business, must not be duplicated in a collection. In the example above the same contact can't be listed twice with the same role on the same project. i.e. together the projectid, contactid, and roleid must be unique. I am merging the entitycollections between projects so I know that the projectid will differ, but I want to know whether there's already an entity in the collection with the same contactid and roleid before adding an entity to the collection. I guess I could add a unique constraint on the table (which I don't currently have) and then just try adding all the entities and catching the exception, but this seems a little odd? Still :

If the starting point is always: check if the contact already performs a given role on a given project, you could use a little trick: create a hashtable and use the contactID as the key. As value you use a 64bit number, with the roleID shifted up 32bits and or-ed with the projectid.

Then traverse the collection once and you can check easily with the hashtable if an entity is already there or not.