Count Customers Who Visited on Multiple Dates

Problem Statement

An Amazon Go store maintains an entry log recording each time a customer walks through the turnstile. Each record consists of a customer_id and a timestamp or date string.

A customer may enter the store multiple times on the same calendar day (e.g., buying a morning coffee and returning for lunch). These same-day visits only count as a single visit date.

Write a function that returns the total number of distinct customers who have visited the store on at least two different calendar dates.

Function Signature

def count_multi_date_visitors(logs: list[list[str]]) -> int:
    pass

Input and Output Format

• Input: A list of pairs logs, where each element is [customer_id, visit_date].
  • customer_id: A string representing the unique identifier for a customer.
  • visit_date: A string representing the date in "YYYY-MM-DD" format.
• Output: An integer representing the count of unique customers who visited on at least two distinct dates.

Constraints

• 0 <= len(logs) <= 2 * 10^5
• 1 <= len(customer_id) <= 32 (alphanumeric characters)
• visit_date is a valid date string in "YYYY-MM-DD" format.

Examples

Example 1

Input: logs = [
    ["cust_1", "2023-11-01"],
    ["cust_2", "2023-11-01"],
    ["cust_1", "2023-11-01"],
    ["cust_1", "2023-11-02"],
    ["cust_3", "2023-11-03"]
]
Output: 1

Explanation:

• cust_1 visited on "2023-11-01" and "2023-11-02" (2 distinct dates).
• cust_2 visited only on "2023-11-01" (1 distinct date).
• cust_3 visited only on "2023-11-03" (1 distinct date). Only cust_1 visited on at least two distinct dates.

Example 2

Input: logs = [
    ["alice", "2023-10-05"],
    ["bob", "2023-10-05"],
    ["alice", "2023-10-06"],
    ["bob", "2023-10-07"],
    ["alice", "2023-10-05"]
]
Output: 2

Explanation: Both alice ("2023-10-05", "2023-10-06") and bob ("2023-10-05", "2023-10-07") visited on two distinct dates.

Edge Cases

• logs is empty: return 0.
• A customer visits multiple times on the same date: multiple entries for the same date must not inflate the distinct date count.
• All visits belong to a single date across all customers: return 0.