GATE CS 1997 Q6.10 — Integrity Constraints

A referential integrity constraint requires that any value of a foreign key in a referencing relation must exist as a primary key value in the referenced relation.

Here, $S$ is the referencing relation (with foreign key $d$) and $R$ is the referenced relation (with primary key, say, $a$). Thus, for every tuple in $S$, the value of $d$ must exist in the primary key column of $R$.

Let's analyze the four operations:

1. I. Insert into $R$: Safe. Adding a new primary key value to $R$ does not affect any existing references in $S$.
2. II. Insert into $S$: Can cause a violation. If we insert a tuple into $S$ with a foreign key value $d$ that does not exist in $R$, the referential integrity is violated.
3. III. Delete from $R$: Can cause a violation. If we delete a tuple from $R$ whose primary key is currently referenced by one or more tuples in $S$, those tuples in $S$ will be left with dangling references.
4. IV. Delete from $S$: Safe. Removing a tuple from the referencing relation $S$ never creates any dangling references.

Therefore, only operations II and III can cause a violation of the referential integrity constraint.