BookMyShow (Ticket Search and Payment)

PayPal's "BookMyShow (Ticket Search and Payment)" interview question focuses on low-level design (LLD) for a movie ticket booking system, tagged for backend, web, and system design roles. It emphasizes implementing core functionalities like cinema management, show scheduling, ticket booking/cancellation, and search features using OOP principles and design patterns.

Problem Statement

Design a movie ticket booking system similar to BookMyShow. Key requirements include:

• Adding new cinema halls and shows within those halls.
• Allowing users to book and cancel tickets.
• Listing all cinemas in a city showing a specific movie.
• Providing a list of all shows in a cinema hall for a particular movie.
• Classes handling search/listing functionalities must update automatically when new shows are added, using the Observer design pattern to notify relevant classes.[1]

Core Classes and Structure

Typical implementation involves these entities (based on standard LLD approaches for this problem):

• Movie: Stores movie ID, name, duration.
• CinemaHall: Manages hall ID, name, total seats, seat layout (e.g., rows/columns).
• Show: Links movie, hall, start time, end time, seat availability (often a 2D matrix or bitset).
• Booking: Tracks user ID, show ID, selected seats, booking status, payment details.
• City/MovieSearchService: Handles queries like "cinemas showing Movie X in City Y"; observes show additions.
• ShowSearchService: Lists shows for a movie in a specific hall; also observer-based.

Observer Pattern Usage: ShowManager notifies CityMovieSearchService and HallMovieSearchService on new show additions to keep search indexes updated.[1]

Key Methods and Input/Output Examples

No official full I/O test cases exist (as it's an LLD whiteboard/OA problem), but standard API signatures from practice implementations include:

| Method | Input | Output | Description | |--------|-------|--------|-------------| | addCinemaHall(hallId, name, rows, cols) | hallId: String, name: String, rows: int, cols: int | boolean (success) | Creates hall with seat matrix. | | addShow(hallId, movieId, startTime, duration) | hallId: String, movieId: String, startTime: long, duration: int | ShowId: String | Adds show; notifies observers. Example: Input ("H1", "M1", 1640000000L, 120) → "S1" [1] | | bookTickets(userId, showId, seatIds[]) | userId: String, showId: String, seatIds: List<String> | Booking or null (if unavailable) | Locks and books seats. Example: Input ("U1", "S1", ["A1", "A2"]) → Booking{bookingId: "B1", status: "CONFIRMED"} [3] | | cancelBooking(bookingId) | bookingId: String | boolean | Frees seats. | | getCinemasByCityMovie(city, movieName) | city: String, movieName: String | List<CinemaHall> | Search example: Input ("Mumbai", "Avengers") → ["PVR1", "INOX2"] [1][3] | | getShowsByHallMovie(hallId, movieName) | hallId: String, movieName: String | List<Show> | Search example: Input ("H1", "Avengers") → [Show{S1: 6PM}, Show{S2: 9PM}] [1] | | blockSeatsTemp(seatIds[], userId) | seatIds: List<String>, userId: String | Timeout: int (e.g., 600s) | Holds seats for payment window. Example from similar: → "Seats blocked for 10 mins" [3] |

Constraints

• Scale: Support 1000+ cinemas, 10K shows/day, 1M concurrent bookings (use efficient seat locking like Redis-inspired TTL or optimistic concurrency).[3]
• Concurrency: FIFO seat allocation; temporary locks (5-10 mins) to prevent double-booking.[3]
• Data: Seats as 2D grid (up to 100x20); times in Unix timestamps; ACID for payments/bookings.
• Edge Cases: Overlapping shows in same hall, cancellations post-payment, city/movie not found (empty lists), invalid seats.[1]

Practice links from sources point to detailed coding: e.g., Codezym Q10 for full implementation.[1]