Description
Homework Description
The data model is the same for this homework as in the previous two homeworks.
Before you start. Add to your solution script the following commands that create copies of two
tables and alters one of them.
create table scores2 as select * from scores ;
create table contestants2 as select * from contestants ;
alter table contestants2 add iswinner boolean ;
Problem 1. We are tired of finding who is the winner of a game, something that does not change
over time. So, it is best to store this directly and use it in queries, instead of recomputing it
every time.
Update the contestants2 table you just created so that iswinner has value True if that
contestant is the winner of the game and False otherwise. To find a winner, find contestants
such that there is no other player with a higher final score. Also, players who do not have a
final score in the database will count as winners.
Note that if two players are tied, they will both count as winners for the purposes of this
problem.
1
Problem 2. We also would like to compute which games were consecutive. We do not know this
for sure because of missing data. However, there are a few things we can compute:
– If two dates are consecutive, then clearly one came after the other.
– If one date is a friday and the other one is the following monday, then they are consecutive.
(You can use extract(isodow from airdate) to find the airdate.
– If two dates d1, d2 are not consecutive according to the above rules, but there is no
other game between these dates and one of the contestants from the game in date d1 is also
competing in date d2, then these games are consecutive. Note that to reduce complexity, we
will only consider games less than 60 days apart.
Based on these rules, create and populate a new relation with the schema:
follows(gameid1, gameid2)
where gameid2 is the id of the game directly following the game given by gameid1.
Problem 3. Now create an after trigger called scores trigger on update of the scores2 table.
Whenever a score is updated, if it is the final round score, then you must update contestants2
table to make sure that the iswinner attributes are correct for the contestants of this game.
To create this trigger, you will need to create a function in the Postgresql syntax. Name this
function scores trigger f.
Problem 4. I guess my parsing was not so good as we are missing a lot of data. To understand
where to start debugging, I would like to find out where my parsing has failed the worst.
To help me, write a single pl/pgsql procedure called:
parsingreport(input year)
that takes as input an integer representing a specific year. Your procedure should return a
report regarding what data is missing in each year:
1. How many contestants have zero responses (to any regular and final clue) in the database
for a game they participated in.
2. How many total clues are missing (including regular and final clues).
Note that each game should have a total of 1+2x6x5= 61 clues. Use this to compute
how many clues there should be.
It is possible that a clue is missing because it was never revealed, but it is much more
likely to be parser error.
3. How many responses to final jeopardy clues are missing
Each contestant with a positive score (above zero) coming into the final jeopardy (i.e.
in Round 3) must have a response for the final clue in the database. Exclude in your
computation any response from a contestant with a missing final score or a negative
score in round 3.
If a final clue is missing, exclude any responses to that clue in your computation as well.
4. Your function should return a single string that reports all of these statistics.
Note that for this question, you can use the original tables as you will be querying them but
not changing them.
Submission Instructions.
Submit a single ASCII text file named username_hw6ans.sql that contains all your queries to
SUBMITTY. It should have the same format as previous homeworks, as shown below.
2
— Print your answer and RCS id first
SELECT ‘Student: Sibel Adali (adalis@rpi.edu)’;
create table scores2 as select * from scores ;
create table contestants2 as select * from contestants ;
alter table contestants2 add iswinner boolean ;
SELECT ‘Problem 1’;
— Insert your solution to Problem 1 here
SELECT ‘Problem 2’;
— Insert your solution to Problem 2 here
— Your solution should only create a table
— follows(gameid1, gameid2)
— and populate it, and nothing else.
SELECT ‘Problem 3’;
— Insert the code for creating your trigger for Problem 3 here.
— Your code should only create a trigger named
— scores_trigger and a function named scores_trigger_f
— and nothing else
SELECT ‘Problem 4’;
— Insert the code for creating your function for Problem 4 here.
— Your code should only create a single function named
— parsingreport(input_year)
— and nothing else
3
Database Schema
— Each game is in a season, given by id
CREATE TABLE games
( id INT — season id
, gameid INT
, airdate DATE
, PRIMARY KEY (gameid)
) ;
— Each contestant is identified by a shortname, which is unique for a
— game.
CREATE TABLE contestants
( gameid INT
, fullname VARCHAR(100)
, description VARCHAR(255)
, shortname VARCHAR(100)
, PRIMARY KEY (gameid, shortname)
, FOREIGN KEY (gameid) REFERENCES games(gameid)
) ;
— The overall scores of each contestants after different rounds
— of the game.
— Rounds ‘1’, ‘2’ are in the first stage called the ‘Jeopardy’ stage,
— Round ‘3’ is after ‘Double Jeopardy’ before ‘Final Jeopardy’.
— Round ‘Final Score’ is the actual score of each person
— Round ‘Coryat Score’ is the hypothetical score without the bets
— Round ‘6’ is an error, which needs to be identified later.
CREATE TABLE scores
( gameid INT
, shortname VARCHAR(100)
, score INT
, round VARCHAR(20)
, PRIMARY KEY (gameid, shortname, round)
, FOREIGN KEY (gameid, shortname)
REFERENCES contestants(gameid, shortname)
) ;
— Each game has many clues, clue is the question, and correct_answer is the answer
— value is the dollar value of the clue: amount player wins/looses
— for correct, incorrect answers
— category is the named of the category
— cat_type is one of: ‘J’: ‘Jeopardy’ round and ‘DJ’: ‘Double Jeopardy’ round
— isdd is true if the question was a double jeopardy question
CREATE TABLE clues
( gameid INT
, clueid INT
, clue TEXT
, value INT
, category VARCHAR(255)
, cat_type VARCHAR(10)
, isdd BOOLEAN
, correct_answer VARCHAR(255)
, PRIMARY KEY (gameid, clueid)
, FOREIGN KEY (gameid) REFERENCES games(gameid)
4
) ;
— Each contestant can answer a clue, if the answer is wrong,
— another contestant can answer. This relation stores all
— contestants who gave a response (but not what they said).
— If there is no correct answer for a question here, it means
— that no contestant answered the question correctly.
CREATE TABLE responses
( gameid INT
, clueid INT
, shortname VARCHAR(255)
, iscorrect BOOLEAN
, PRIMARY KEY (gameid, clueid, shortname)
, FOREIGN KEY (gameid, clueid) REFERENCES clues(gameid, clueid)
, FOREIGN KEY (gameid, shortname)
REFERENCES contestants(gameid, shortname)
) ;
— At the end of the game, there is a single question/clue called
— the ‘Final Jeopardy’. This relation stores the clues for this
— specific round. There is no dollar value attached to these questions.
CREATE TABLE final_clues
( gameid INT
, clue TEXT
, category VARCHAR(255)
, correct_answer VARCHAR(255)
, PRIMARY KEY (gameid)
, FOREIGN KEY (gameid) REFERENCES games(gameid)
) ;
— For the ‘final jeopardy’, all contestants give an answer and a bet
— The bet is the dollar amount the contestant will win/loose if they
— answer correctly. Only contestants with positive winnings/scores
— at round ‘3’ can participate. This relation stores the bets and
— whether each person scored correctly or not.
CREATE TABLE final_responses
( gameid INT
, shortname VARCHAR(255)
, iscorrect BOOLEAN
, bet FLOAT — VARCHAR(10)
, PRIMARY KEY (gameid, shortname)
, FOREIGN KEY (gameid, shortname)
REFERENCES contestants(gameid, shortname)
) ;
CREATE TABLE states
( name VARCHAR(50)
, population INT
) ;
5
