Untitled

Query for exact match of users in a conversation in SQL Server
CONVERSATION
Id, Subject, Type

USERCONVERSATION
Id, UserId, ConversationId

SELECT ConversationID
FROM UserConversation
GROUP BY ConversationID
HAVING
   Count(UserID) = 3 -- this isn't necessary but might improve performance
   AND Sum(CASE WHEN UserID IN (1, 2, 3) THEN 1 ELSE 0 END) = 3

SELECT ConversationID
FROM
   UserConversation UC
   LEFT JOIN (
      SELECT 1 UNION ALL SELECT 2 UNION ALL SELECT 3
   ) U (UserID) ON UC.UserID = U.UserID
GROUP BY ConversationID
HAVING
   Count(U.UserID) = 3
   AND Count(UC.UserID) = 3

SELECT
   Coalesce(C.ConversationID, UC.ConversationID) ConversationID
   -- Or could be Min(C.ConversationID)
FROM
   Conversation C
   CROSS JOIN (
      SELECT 1 UNION ALL SELECT 2 UNION ALL SELECT 3
   ) U (UserID)
   FULL JOIN UserConversation UC
      ON C.ConversationID = UC.ConversationID
      AND U.UserID = UC.UserID
GROUP BY Coalesce(C.ConversationID, UC.ConversationID)
HAVING Count(*) = Count(U.UserID)