CODEWITHSUNDEEP
sqldatabaseprologalgebrarelational
pmax1
pmax1

Reputation: 236

Prolog Database Duplicates

I'm new to Prolog. My "database" consists of the following rules: 

teacher(1, 'Benjamin', 'Johnson').
teacher(2, 'Mike', 'Oliver').
teacher(3, 'James', 'Johnson').

school(1, 'School1','Town1').
school(2, 'School2','Town2').
school(3, 'School3','Town3').

group(1,'Group1', 1, 1).
group(2,'Group2', 3, 2).
group(3,'Group3', 2, 1).
group(4,'Group4', 2, 1).
group(5,'Group5', 3, 3).
group(6,'Group6', 1, 3).

"teacher" has the scheme:

teacher(TeacherID, TeacherFirstname, TeacherLastname)

"school" has the scheme: 

school(SchoolID, SchoolName, City)

"group" has the scheme: 

group(GroupID, GroupName, TeacherID, SchoolID)

I want to output all schools where teachers work whose last name is 'Johnson'.

My query is as follows: 

group( GroupID, _ , TeacherID , SchoolID )  , school( SchoolID , SchoolName, _ ) , teacher( TeacherID , _ , 'Johnson' ).

I want to return the name of the schools, who fit the query, only once. However using my current query I get a single school multiple times. How can I detect duplicates and not output them? Thanks!

PS: The relational algebra would be

 π SchoolName (σ TeacherLastname = 'Johnson' (teacher ⨝ group ⨝ school))

Upvotes: 0

Views: 115

Answers (1)

Isabelle Newbie
Isabelle Newbie

Reputation: 9378

With an auxiliary predicate:

school_with_teacher_johnson(School) :-
    group(_, _, Teacher, SchoolID),
    school(SchoolID, School, _),
    teacher(Teacher, _, 'Johnson').

You can use the setof/3 predicate with gives you a set of solutions, that is, a sorted list with duplicates removed:

?- setof(School, school_with_teacher_johnson(School), Schools).
Schools = ['School1', 'School2', 'School3'].

It's also possible to use setof/3 without an auxiliary predicate, but it's ugly and easy to get wrong.

Upvotes: 2

Related Questions