User Pairs with Peak Shared Followers

Medium

A social intelligence platform stores directed follow edges and needs to identify the strongest audience-overlap relationships between creators. Table: Relations user_id: creator being followed follower_id: account following that creator Data guarantees: (user_id, follower_id) is unique. Each row represents one active follow edge. Task: For every ordered-free pair of distinct creators, count how many followers they have in common. Return all creator pairs tied for the largest shared-follower count. Output requirements: Return exactly two columns: user1_id user2_id Keep only rows where user1_id < user2_id. Include every pair tied for the global maximum. Output order may be arbitrary. Supported submission environments: SQL (SQLite 3.27.2) Pandas (Python)

Example 1

Input

Relations:
| user_id | follower_id |
|---------|-------------|
| 1       | 3           |
| 2       | 3           |
| 7       | 3           |
| 1       | 4           |
| 2       | 4           |
| 7       | 4           |
| 1       | 5           |
| 2       | 6           |
| 7       | 5           |

Output

[
  {"user1_id":1,"user2_id":7}
]

Explanation

Pair (1, 7) shares followers {3,4,5} (count 3), while (1,2) and (2,7) each share only {3,4} (count 2).

Example 2

Input

Relations:
| user_id | follower_id |
|---------|-------------|
| 10      | 100         |
| 20      | 100         |
| 10      | 101         |
| 20      | 101         |
| 30      | 102         |
| 40      | 102         |
| 30      | 103         |
| 40      | 103         |

Output

[
  {"user1_id":10,"user2_id":20},
  {"user1_id":30,"user2_id":40}
]

Explanation

Both pairs have a shared-follower count of 2, which is the global maximum, so both must be returned.

Example 3

Input

Relations:
| user_id | follower_id |
|---------|-------------|
| 1       | 1000        |
| 2       | 2000        |
| 3       | 3000        |

Output

[]

Explanation

No follower appears under two different users, so no user pair has a positive common-follower count.

Constraints

0 <= number of rows in Relations <= 300000
0 <= user_id <= 1000000000
0 <= follower_id <= 1000000000
(user_id, follower_id) is unique
Return exactly columns: user1_id, user2_id
Only pairs with user1_id < user2_id are valid
Return all pairs tied for the global maximum shared-follower count
If no pair has at least one common follower, return an empty result
For SQL submissions, use SQLite 3.27.2 semantics
For Pandas submissions, return a DataFrame with columns: user1_id, user2_id

Relations: | user_id | follower_id | |---------|-------------| | 1 | 3 | | 2 | 3 | | 7 | 3 | | 1 | 4 | | 2 | 4 | | 7 | 4 | | 1 | 5 | | 2 | 6 | | 7 | 5 |

Relations: | user_id | follower_id | |---------|-------------| | 10 | 100 | | 20 | 100 | | 10 | 101 | | 20 | 101 | | 30 | 102 | | 40 | 102 | | 30 | 103 | | 40 | 103 |