Description
In this homework, you will be working with National Basketball Association (NBA) player data. In the
Modules section on Canvas, under DATA, you will find two files: README and player-stats.sql. The former
defines the table columns that appear in the latter. It is provided for completeness; you should not have to
refer to it.
Question 1
(10 points)
Download player-stats.sql and read it into your postgres session using the \i command. Show that
you have changed the working directory inside postgres to where your downloaded data are stored. As a
final step, display the current list of tables in your database.
FILL ME IN
1
Question 2
(10 points)
Use an aggregate function from Lecture 8 to determine how many unique colleges there are among the
476 players in the database. (You might want to look up the documentation for distinct, which is SQL’s
analogue to R’s unique().)
FILL ME IN
Question 3
(10 points)
Create a table called teams, which will have five columns: an id column of type char(3) that will be
the table’s primary key; location and name columns of type text (e.g., Boston for location, Celtics for
name); a column showing the division in which the team plays (e.g., Atlantic); and a column showing
the conference the team plays in (e.g., Eastern). Note that division and conference should be of type
DivisionType and ConferenceType respectively; these types should be created as enumerated variables.
(See the postgres documentation for CREATE TYPE and look specifically for the variant that contains the
word ENUM; essentially, you are defining your own factor variables, with defined levels.)
FILL ME IN
Run the following code in your postgres() session to populate the teams table:
insert into teams VALUES
(’BOS’, ’Boston’, ’Celtics’, ’Atlantic’, ’Eastern’),
(’BKN’, ’Brooklyn’, ’Nets’, ’Atlantic’, ’Eastern’),
(’NYK’, ’New York’, ’Knicks’, ’Atlantic’, ’Eastern’),
(’PHI’, ’Philadelphia’, ’76ers’, ’Atlantic’, ’Eastern’),
(’TOR’, ’Toronto’, ’Raptors’, ’Atlantic’, ’Eastern’),
(’CHI’, ’Chicago’, ’Bulls’, ’Central’, ’Eastern’),
(’CLE’, ’Cleveland’, ’Cavaliers’, ’Central’, ’Eastern’),
(’DET’, ’Detroit’, ’Pistons’, ’Central’, ’Eastern’),
(’IND’, ’Indiana’, ’Pacers’, ’Central’, ’Eastern’),
(’MIL’, ’Milwaukee’, ’Bucks’, ’Central’, ’Eastern’),
(’ATL’, ’Atlanta’, ’Hawks’, ’Southeast’, ’Eastern’),
(’CHA’, ’Charlotte’, ’Bobcats’, ’Southeast’, ’Eastern’),
(’MIA’, ’Miami’, ’Heat’, ’Southeast’, ’Eastern’),
(’ORL’, ’Orlando’, ’Magic’, ’Southeast’, ’Eastern’),
(’WAS’, ’Washington’, ’Wizards’, ’Southeast’, ’Eastern’),
(’DEN’, ’Denver’, ’Nuggets’, ’Northwest’, ’Western’),
(’MIN’, ’Minnesota’, ’Timberwolves’, ’Northwest’, ’Western’),
(’OKC’, ’Oklahoma City’, ’Thunder’, ’Northwest’, ’Western’),
(’POR’, ’Portland’, ’Trail Blazers’, ’Northwest’, ’Western’),
(’UTA’, ’Utah’, ’Jazz’, ’Northwest’, ’Western’),
(’GSW’, ’Golden State’, ’Warriors’, ’Pacific’, ’Western’),
(’LAC’, ’Los Angeles’, ’Clippers’, ’Pacific’, ’Western’),
(’LAL’, ’Los Angeles’, ’Lakers’, ’Pacific’, ’Western’),
(’PHX’, ’Phoenix’, ’Suns’, ’Pacific’, ’Western’),
2
(’SAC’, ’Sacramento’, ’Kings’, ’Pacific’, ’Western’),
(’DAL’, ’Dallas’, ’Mavericks’, ’Southwest’, ’Western’),
(’HOU’, ’Houston’, ’Rockets’, ’Southwest’, ’Western’),
(’MEM’, ’Memphis’, ’Grizzlies’, ’Southwest’, ’Western’),
(’NOP’, ’New Orleans’, ’Hornets’, ’Southwest’, ’Western’),
(’SAS’, ’San Antonio’, ’Spurs’, ’Southwest’, ’Western’);
Question 4
(20 points)
None of the available datasets list player positions. In the NBA, there are five commonly acknowledged
positions: center, power forward, small forward, shooting guard, and point guard. One can map information
in the more_player_stats table to these positions via the empirical formula
prl = per – 67*va/(gp*minutes)
where ranges of prl map directly to positions (see below). Create a new table (just call it new_table) that
has three columns: player, of type integer references more_player_stats (it’s a foreign key, something
we haven’t explicitly mentioned yet, but is useful for joining tables as we will see next week); prl, of type
numeric; and position, of type text. Use insert to populate this table: you’d select the id from
more_player_stats, and you’d select the equation above. That sounds a bit weird. Let’s say table foo has
three columns, id, x, and y. Now you create bar, that has new_id and z, which is defined as x/y.
• insert into bar (new_id,z) (select id,round(x/y,1) from foo);
(Here I decided to round off the quotient. You should do the same with prl.) The last thing to do is to
update your new table by setting the position on the basis of the following criteria: ‘PF’ where prl is greater
than or equal to 11.3; ‘PG’ where prl is [10.8,11.3); ‘C’ for [10.6,10.8); and ‘SG/SF’ for [0,10.6). Display
the first 10 rows of your new table.
FILL ME IN
Question 5
(20 points)
Take the position column you’ve just defined and add it to the player_bios table. (Remember: alter
table and update.)
You’ll take advantage of the fact that player in new_table references id in more_player_stats. . . which
means it references id in player_bios as well. After you are done, display the first five rows of player_bios
(selecting just the firstname, lastname, and position columns).
FILL ME IN
Question 6
(20 points)
Now we will convert the heights given in player_bios from feet-inches format to simply inches. This involves
altering the player_bios table to add a new column of type numeric, then using update to set the values
of the height in inches. You may wish to use the following:
3
12*split_part(height,’-‘,1)::integer + split_part(height,’-‘,2)::integer
This splits the height string on ‘-’; the first output is feet, so we multiply that by 12 (and cast to integer),
and the second output is inches, which is simply cast to integer. Finally, use two alter commands to drop
the old height column and to rename your new column ‘height’. As you did in the last question, display the
first five rows of player_bios (but this time: firstname, lastname, and height).
FILL ME IN
4
