SQL and Relational Theory, 2nd Edition

EXAMPLE 7: DEALING WITH AMBIGUITY

As we saw in Chapter 10, natural language is often ambiguous. For example, consider the following query: “Get suppliers such that every part they supply is in the same city.” First of all, notice the subtle (?) difference between this example and the previous one. Second, and more important, note that this natural language formulation is indeed ambiguous! For the sake of definiteness, I’m going to assume it means the following:

Get suppliers SX such that for all parts PX and PY, if SX supplies both of them, then PX.CITY = PY.CITY.

Observe that a supplier who supplies just one part will qualify under this interpretation. (So will a supplier who supplies no parts at all, incidentally.) Alternatively, the query might mean:

Get suppliers SX such that (a) SX supplies at least two distinct parts and (b) for all pairs of distinct parts PX and PY, if SX supplies both of them, then PX.CITY = PY.CITY.

Now a supplier who supplies just one part or no parts at all won’t qualify.

As I’ve said, I’m going to assume the first interpretation, just to be definite. But note that ambiguities of this kind are quite common with complex queries and complex business rules, and another advantage of logic in the context at hand is precisely that it can pinpoint and help resolve such ambiguities.

Here then is a logical formulation for the first interpretation:

 { SX } WHERE FORALL PX ( FORALL PY ( IF EXISTS SPX ( SPX.SNO = SX.SNO AND SPX.PNO = PX.PNO ) AND EXISTS SPY ( SPY.SNO = ...