数据库设计-我的数据应该有多精细？

Question

Suppose I have an application that renders a seating chart to the user.

There are three seating sections.

Each section contains five rows.

Each row contains a variable number of seats.

Each seat has its own attributes, among which is an association with the customer who has purchased the seat.

Would it make sense for me to model the data into the following tables?

• floor_plan
• seating_section
• seating_row
• seat
• customer

Ultimately, this data will need to be aggregated so that it is meaningful and useful to my front end. How does one go about assembling data from the database into something that is useful for and specific to a given view?

Also, I have a gazillion more questions regarding similar database-design-related items. Are there any good books that could give me a solid foundation in this stuff?

Answer 1

From the relational point of view, data should be granular enough that

• the dbms can enforce sensible constraints on it, and
• client code never has to parse it.

Assuming there's only one venue (only one floorplan), the enterprise will typically identify a seat by its section, row, and number. Principles are the same for multiple floorplans and multiple venues.

Let's assume that section 1 has 3 rows, section 2 has 5 rows, and section 3 has 4 rows. (Tested in PostgreSQL.)

create table sections (
  section_num integer primary key
    check (section_num between 1 and 3)
);

insert into sections values (1), (2), (3);

create table section_rows (
  section_num integer not null
    references sections (section_num),
  row_num integer not null
    check (row_num between 1 and 5),
  primary key (section_num, row_num)
);

insert into section_rows values
(1,1), (1,2), (1,3),
(2,1), (2,2), (2,3), (2,4), (2,5),
(3,1), (3,2), (3,3), (3,4);

create table seats (
  section_num integer not null,
  row_num integer not null,
  seat_num integer not null,
  primary key (section_num, row_num, seat_num),
  foreign key (section_num, row_num) 
    references section_rows (section_num, row_num)
);

insert into seats values
(1, 1, 1), (1, 1, 2), (1, 1, 3),
(1, 2, 1), (1, 2, 2), (1, 2, 3),
(1, 3, 1), (1, 3, 2), (1, 3, 3), (1, 3, 4),
(2, 1, 1), (2, 1, 2), (2, 1, 3),
(2, 2, 1), (2, 2, 2), (2, 2, 3),
(2, 3, 1), (2, 3, 2), (2, 3, 3), (2, 3, 4),
(2, 4, 1), (2, 4, 2), (2, 4, 3), (2, 4, 4),
(2, 5, 1), (2, 5, 2), (2, 5, 3), (2, 5, 4), (2, 5, 5),
(3, 1, 1), (3, 1, 2), (3, 1, 3),
(3, 2, 1), (3, 2, 2), (3, 2, 3),
(3, 3, 1), (3, 3, 2), (3, 3, 3), (3, 3, 4),
(3, 4, 1), (3, 4, 2), (3, 4, 3), (3, 4, 4);

This last table, "seats" identifies every seat in the venue. Once these three tables are populated, you never have to change them unless you tear out seats or install new ones.

Now you can sell each one to a customer.

create table event_sales (
  -- Assumes an event identifier identifies the date and time as well
  -- as the event's name.
  event_id integer not null, -- references events (not shown)
  section_num integer not null,
  row_num integer not null,
  seat_num integer not null,
  customer_columns_go_here char(1) default 'x',
  primary key (event_id, section_num, row_num, seat_num),
  foreign key (section_num, row_num, seat_num) 
    references seats (section_num, row_num, seat_num)
);

insert into event_sales values
(1, 1, 1, 1, 'a'),
(1, 1, 1, 2, 'a'),
(1, 1, 1, 3, 'a'),
(1, 2, 2, 1, 'b'),
(2, 2, 1, 1, 'a'),
(2, 2, 1, 2, 'b'),
(2, 2, 1, 3, 'c'),
(2, 3, 2, 1, 'd');

All these tables are in at least 5NF.

What seats are available for event 1? (Probably the most common query for a seating application.)

select *
from seats
except
(select section_num, row_num, seat_num from event_sales where event_id = 1)
order by section_num, row_num, seat_num;

Database design is a bigger topic than most people think it is. You're not likely to get it right by skimming a couple of web sites. Avoid bad practices while you learn. I think you're probably best served by Bill Karwin's book SQL Antipatterns: Avoiding the Pitfalls of Database Programming .