7.%20Refinement%20Queries

1
7. Refinement Queries
Three different assumptions about the input
database
Tuple Inside Tuple Outside
Closed World True False
Open World True May be
Possible World May be False
2

• UOR Universal Object Relation
• all attributes in one constraint relation
• assumes possible world

Example
Alfred_Age ... Harold_Age
Alfred.y Harold.y Alfred.y lt 70 Gerald.y 0 Harold.y 5 Bernard.y Edward.y gt 25 Elise.y Fred.y gt 4
3

• 7.3 Syntax
• Refinement rule body same as Datalog rule body
• Refinement rule head a constraint
• Example
• x.y lt z.y Sibling(x, z).
• z.y x.y Parent(x, z).
• x.y ? 0 Person(x).

4

• Example
• Suppose we have five sets of persons
• P passed drivers test
• L licensed person
• A allowed to drive
• D drives sometime
• I intoxicated / drinks alcohol

5

• We know the following
• All(L, P)
• All persons must pass a test before getting a
  drivers license.
• All(A,L)
• All persons allowed to drive must have a
  license.
• All(D,P)
• No one drives without having passed the test.
• All(L,I)
• No one who drinks has a license.
• Some(D,I)
• Some person X both drives and drinks.

6

• Here we have UOR with scheme
• UOR(A.S, D.S, I.S, L.S, P.S)
• x.S ? y.S All(x,y)
• x.S ? y.S ? ? Some(x,y)

7

• 7.5 Projection Queries
• SELECT y1,.yk
• FROM UOR
• Example
• SELECT Donald_Age
• FROM UOR
• This will return a table with the attribute
  Donald_Age
• 47 Donald.y , Donald.y 49