Weighted Election Results from Split Ballots

Medium

A city analytics platform records digital ballots where a voter may support multiple candidates. Table: Votes voter: voter identifier candidate: candidate name (NULL means this voter abstained) Weighted voting policy: Every voter has total vote power equal to 1. If a voter selects k non-null candidates, each selected candidate receives 1/k from that voter. A NULL candidate row contributes 0 and represents abstention. Task: Compute total weighted votes per candidate and return every candidate whose total is maximum. Output requirements: Return exactly one column in this order: candidate If multiple candidates tie for first place, return all of them. Sort output by candidate ASC. Supported submission environments: SQL (SQLite 3.27.2) Pandas (Python)

Example 1

Input

Votes:
| voter    | candidate  |
|----------|------------|
| Kathy    | null       |
| Charles  | Ryan       |
| Charles  | Christine  |
| Charles  | Kathy      |
| Benjamin | Christine  |
| Anthony  | Ryan       |
| Edward   | Ryan       |
| Terry    | null       |
| Evelyn   | Kathy      |
| Arthur   | Christine  |

Output

[
  {"candidate":"Christine"},
  {"candidate":"Ryan"}
]

Explanation

Charles splits one vote equally across three candidates (1/3 each). Kathy and Terry abstain (0 contribution). Christine and Ryan both finish with 2.333... weighted votes, so both are winners.

Example 2

Input

Votes:
| voter | candidate |
|-------|-----------|
| V1    | A         |
| V1    | B         |
| V2    | A         |
| V3    | A         |
| V3    | B         |
| V3    | C         |
| V4    | C         |

Output

[
  {"candidate":"A"}
]

Explanation

A receives 1/2 + 1 + 1/3 = 1.833..., B receives 1/2 + 1/3 = 0.833..., and C receives 1/3 + 1 = 1.333.... A is the unique winner.

Example 3

Input

Votes:
| voter | candidate |
|-------|-----------|
| U1    | Delta     |
| U1    | Alpha     |
| U2    | Alpha     |
| U3    | Delta     |
| U4    | null      |

Output

[
  {"candidate":"Alpha"},
  {"candidate":"Delta"}
]

Explanation

Alpha gets 0.5 + 1 = 1.5 and Delta gets 0.5 + 1 = 1.5. The abstaining voter contributes nothing, so both tied winners are returned in ascending order.

Constraints

1 <= number of rows in Votes <= 200000
voter is a non-empty string
candidate is either NULL (abstain) or a non-empty string
(voter, candidate) is unique
Each voter either abstains once (candidate=NULL) or votes for one or more non-null candidates
At least one non-null candidate vote exists in each test case
Return exactly one column: candidate
Return all co-winners if weighted totals tie at maximum
Sort output by candidate ASC
For SQL submissions, use SQLite 3.27.2 semantics
For Pandas submissions, return a DataFrame with exact column [candidate]

Votes: | voter | candidate | |----------|------------| | Kathy | null | | Charles | Ryan | | Charles | Christine | | Charles | Kathy | | Benjamin | Christine | | Anthony | Ryan | | Edward | Ryan | | Terry | null | | Evelyn | Kathy | | Arthur | Christine |

Votes: | voter | candidate | |-------|-----------| | V1 | A | | V1 | B | | V2 | A | | V3 | A | | V3 | B | | V3 | C | | V4 | C |