Direct Connections Without Shared Contacts

A social network analytics team is auditing direct connection quality.

Table: Friends

• user_id1: first user id
• user_id2: second user id
• each row means these two users are directly connected

A direct pair (u, v) should appear in the output only when they do not have any shared mutual contact. In other words, there must be no third user w such that:

• (u, w) is a friendship, and
• (v, w) is also a friendship.

Return all qualifying direct pairs.

Output requirements:

• Return exactly these columns: user_id1, user_id2
• Sort result by user_id1 ascending, then user_id2 ascending

Notes:

• Treat friendships as undirected relationships.
• Input pairs are unique.

Supported submission environments:

• SQL (SQLite 3.27.2)
• Pandas (Python)

Friends:
| user_id1 | user_id2 |
|----------|----------|
| 1        | 2        |
| 2        | 3        |
| 2        | 4        |
| 1        | 5        |
| 6        | 7        |
| 3        | 4        |
| 2        | 5        |
| 8        | 9        |

[
  {"user_id1":6,"user_id2":7},
  {"user_id1":8,"user_id2":9}
]

Friends:
| user_id1 | user_id2 |
|----------|----------|
| 10       | 11       |
| 11       | 12       |
| 12       | 13       |

[
  {"user_id1":10,"user_id2":11},
  {"user_id1":11,"user_id2":12},
  {"user_id1":12,"user_id2":13}
]

Friends:
| user_id1 | user_id2 |
|----------|----------|
| 21       | 22       |
| 22       | 23       |
| 21       | 23       |

[]