DeAL TutorialBuilt-In Aggregates

Built-In Aggregates (.deal | .fac)

Given the following facts:

father(joe, peter).
father(joe, mary).
father(jim, lucy).
father(jim, arnold).
father(jim, jim).

Then to list the number of children for each father we can use a Built-In count as follows: 

query no_of_children(X, No).
no_of_children(X, count<Y>) <- father(X, Y).

This will return:

no_of_children(joe, 2)
no_of_children(jim, 3)

Thus, a rule with a Built-In aggregate applies the aggregate (count in the example) to every true body, grouped by the non-aggregate arguments in the head (X in the example).

DeAL supports SQL's five basic aggregates: sum, count, max, min, avg. The count distinct equivalent is also supported.

As discussed later, DeAL supports user-defined aggregates.

Consider another example:

salary(john, 50000, 'Accounting').
salary(james, 60000, 'Accounting').
salary(becky, 100000, 'IT').
salary(laura, 70000, 'Accounting').
salary(sam, 60000, 'IT').

To find the average salary of employees by department we use the following rule: 

query avg_salary(Dept, Sal).
avg_salary(Dept, avg<S>) <- salary(_, S, Dept).

This will return:

avg_salary('Accounting', 60000)
avg_salary('IT', 80000)
Built-in Aggregates in DeAL
Name Semantics SQL Equivalent
sum Compute the sum of items in column SUM
avg Compute the average of items in column AVG
count Count the items in column COUNT
max Find the maximum value in column MAX
min Find the minimum value in column MIN
countd Count the number of distinct items in column COUNT DISTINCT

Multiple Aggregates in the Head

Multiple aggregates are allowed in the same head.

p(K1,K2,...,Km, aggr1<A1>, aggr2<A2>, ..., aggrN<An>) <- Rule Body.

The following conventions apply:

  1. The arguments in the head, that is, K1, ...,Km and A1,...,An must all appear in the body of the rule.
  2. aggr1,...,aggrN can either be builtin aggregates or user-defined aggregates.
  3. Each aggr1<A1>, aggr2<A2>, ..., aggrN<An> is grouped by K1,K2,...,Km where m denote a non-negative integer (thus an empty group-by list is also allowed).
  4. The cartesian product of the results of aggr1, ...,aggrN will be returned for each new value X. Thus, if any of the N aggregates fail for a given X no value is returned at that point.

Duplicates

The DeAL System uses the following policy for duplicate elimination:

  • Duplication in base relations is controlled by the keys of the relations
  • Duplicates are eliminated from the tuples returned as query answers
  • Duplicates are eliminated during the computation of recursive predicates
  • For a head aggregate such as p(X, count<Y>) <- ... the first argument X is a unique key for the values produced by this rule. Similar rules apply to all other Built-In aggregates.
Last Updated 2014