Question

I have a table ALPHA with 2 fields GroupId,Member:

GroupId | Member;
A1----------A;
A1----------B; 
A1----------C;
A2----------A;
A2----------B;
A3----------A;
A3----------D;
A3----------E;

Objective: Given the input of – A,B,C – I have to query the table to find if a GroupId exists for this exact set of members. So, this is what I plan to do:

1. Query the table for all GroupIds whose count is 3 (since my inpt is A,B,C ..I knw its 3)
2. This will give me A1,A3. Now, query this set for exact matching Member values..which will give me A1.

I plan to write a Stored Procedure and would achieve the objective somehow. But, my question can this be achieved in a single query…a single self-join perhaps.

Clarification: The set of (A,B,C) is unique to A1. And if give an input of (A,B,C,D) the query should NOT return A1.

Answer 1

Answers given so far assume that the Member field is unique for any given GroupID. In work I have done this isn’t the case. And also if the group has what you’re looking for, plus some extra, you need to exlucde that group.

SELECT
   [Alpha].GroupID
FROM
   [Alpha]
GROUP BY
   [Alpha].GroupID
HAVING
       SUM(CASE WHEN [alpha].Member IN ('A','B','C') THEN 1 ELSE 0 END) = 3
   AND MIN(CASE WHEN [alpha].Member IN ('A','B','C') THEN 1 ELSE 0 END) = 1

You can also replace the IN clause with a join on to a table holding the members you are searching for…

SELECT
   [Alpha].GroupID
FROM
   [Alpha]
LEFT JOIN
   [Search]
       ON [Search].Member
GROUP BY
   [Alpha].GroupID
HAVING
       SUM(CASE WHEN [alpha].Member = [search].Member THEN 1 ELSE 0 END) = (SELECT COUNT(*) FROM [search])
   AND MIN(CASE WHEN [alpha].Member = [search].Member THEN 1 ELSE 0 END) = 1