Question

I have a table for Projects, a table for Documents (with a FK to Projects), and a table for users with a relationship to projects and documents. Users can be subscribers to a document and that is how they have a relationship to the document.

Users are considered team members on a project. A document can only assigned to one project at a time. Only users, from the “Project Team” can be subscribed to a document.

What I am struggling with, in my application, is that admins are able to add users as subscribers to a document. However, the admin can only pick from a list of unsubscribed users. Those unsubscribed users are members of the project team, like I said above.

Right now, I am creating 2 queries on my database to pull all the subscribed users and the entire list of team members. I then compare the list and only pull the team members that are not subscribed.

I am not sure if I should even index this type of data or just pull from the database directly. If I should use an index, this data needs to be updated quickly becuase the admins need that unsubscribed list rather fast.

This is what my query looks like going against Entity Framework 4.1:

var currentSubscribers = _subscriptionRepository.Get(s => s.DocumentId == documentId).Select(s => s.Contact).ToList();

                if (projecTeamMembers != null)
                {
                    var availableSubscribers = (projecTeamMembers.Where(
                        p => !(currentSubscribers.Select(c => c.Id)).Contains(p.Id))).ToDictionary(c => c.Id, c=> c.LastName + ", " + c.FirstName);
                    return availableSubscribers;
                }
                else
                {
                    return null;
                }

This works great in EF, but I have been thinking of indexing my data using Lucene.Net and need some advice on if I should do this or not.

Answer 1

The best way I have found to do something like this is, using Lucene.net, is to keep an index of subscribers and an index of all team members. And compare the two. It’s faster than pulling form the database each time.