Abstract
The query equivalence problem has been studied extensively for set-semantics and, more recently, for bag and bag-set semantics. However, SQL queries often combine set, bag and bag-set semantics. For example, an SQL query that returns a multiset of elements may call a subquery or view that returns a set of elements. Queries may access both relations that do not contain duplicates, as well as relations with duplicates. As another example, in SQL one can compute a multiset-union of queries, each of which returns a set of answers. This paper presents combined semantics, which formally models query evaluation combining set, bag and bag-set semantics. The equivalence problem for queries evaluated under combined semantics is studied. A sufficient condition for equivalence is presented. For several important common classes of queries necessary and sufficient conditions for equivalence are presented.
| Original language | English |
|---|---|
| Pages (from-to) | 765-785 |
| Number of pages | 21 |
| Journal | VLDB Journal |
| Volume | 18 |
| Issue number | 3 |
| DOIs | |
| State | Published - Jun 2009 |
Bibliographical note
Funding Information:An early version of this article appeared in [7]. This paper extends [7] to include bag semantics, in addition to set and bag-set semantics. This work was partially supported by the Israel Science Foundation (Grant 1032/05).
Keywords
- Bag semantics
- Combined semantics
- Datalog
- Query equivalence
- Set semantics