|
|
Line 4: |
Line 4: |
| * $Q(u,v) \leftarrow T(w,u), U(v,w), R(v,u)$ | | * $Q(u,v) \leftarrow T(w,u), U(v,w), R(v,u)$ |
| * For this query, [[First Order Logic|FOL]] meaning is | | * For this query, [[First Order Logic|FOL]] meaning is |
− | * $\forall \ u, v \Big[ Q(u, v) \Rightarrow \exists \ w \ : \ T(w, u) \land U(v, w) \land R(v, u) \Big]$ | + | * $\forall \ u, v \ \Big[ Q(u, v) \Rightarrow \exists \ w \ : \ T(w, u) \land U(v, w) \land R(v, u) \Big]$ |
| | | |
| Now consider that we have a tuple $(a, b)$ | | Now consider that we have a tuple $(a, b)$ |
Line 29: |
Line 29: |
| | | |
| == Sources == | | == Sources == |
− | * Web Data Management book [http://webdam.inria.fr/Jorge] | + | * [[Web Data Management (book)]] |
| | | |
| [[Category:Logic]] | | [[Category:Logic]] |
Latest revision as of 15:39, 23 November 2015
Skolem Function
Intuition
Suppose, we have the following Conjunctive Query:
- $Q(u,v) \leftarrow T(w,u), U(v,w), R(v,u)$
- For this query, FOL meaning is
- $\forall \ u, v \ \Big[ Q(u, v) \Rightarrow \exists \ w \ : \ T(w, u) \land U(v, w) \land R(v, u) \Big]$
Now consider that we have a tuple $(a, b)$
- $(a, b)$ belongs to the data source that backs $V_1$
- so we have a fact $Q(a, b)$
- from this fact $Q(a, b)$ can infer that $R(b, a)$
- $Q(a, b) \Rightarrow R(b, a)$
- (all conjuncts have to be true for a statement to be true, so it means the last conjuncts holds true)
But we can infer other things as well
- e.g. $Q(a, b) \Rightarrow \exists \ d_1 \ : \ T(d_1, a) \land U(b, d_1)$
- where $d_1$ is some constant
- we don't know its value, but we know it exists (since it's existentially qualified) and
- it depends on constants $a$ and $b$
- so we can denote this dependency as $d_1 = f_1(a, b)$
Skolem Function
- the symbol $f_1(u, v)$ is a Skolem Function of arity 2
- $f_1(u, v)$ denotes that there exists some constant that depends on values of $u$ and $v$
- given two distinct Skolem terms, e.g. $f_1(1, 2)$ and $f_1(2, v_3)$ we never can say if they belong to the same constant or not
Sources