Question:
AirwaysCentre, an international airline booking company currently uses a relational database system for their operations, which includes the following relations:
PASSENGER(p_no, p_name, p_address, p_phone)
FLIGHT(flight_code, airline_name, from_airport, to_airport, depart_time, arrive_time)
BOOKING(booking_no, p_no, invoice_total, payment_made)
FLIGHT_BOOKING(booking_no, flight_code, flight_date, price)
The following assumptions are made:
- The flight code is unique, which includes information on an airline and its flight number, for example „BA300?
- For each flight code, there is only one associated flight on any particular day
- A passenger may purchase tickets for several different flights under one booking number
- However, a passenger may not purchase more than one ticket for the same flight under the same booking number
- The PASSENGER relation contains 2000 records
- The FLIGHT relation contains 500 records
- The BOOKING relation contains 5000 records
- The FLIGHT-BOOKING relation contains 10000 records
- Each flight has a maximum of 200 bookable tickets (seats)
(a) Query optimization in database system aims to improve efficiency in query evaluation. One of the stages involved in the query optimization is to convert some internal form of the query representation into its equivalent but more efficient canonical form using the transformation rules.
Demonstrate, by using two query examples based on the given relations, how appropriate transformation rules can be applied to carry out such a conversion.
(b) Consider the following query based on the above relations and formulate it using correct SQL syntax:
"Find the names and numbers of all passenger who have booked the Air Mauritius flight MK356 for 2nd December 2010".
(c) Specify two possible options to evaluate the query in (b)above. Express each of these options firstly in relational algebra and secondly in a query tree. Clearly show all the steps required in the evaluation of the query.
(d) For each option identified above, calculate the size (number of records) for each of the immediate relations to be produced at each steps. Recommend the best option based on the results of your calculations.