Auto-completion of Underspecied SQL Queries

Formulating SQL queries involving joins is tedious, error-prone, and requires in-depth schema knowledge. We demonstrate a modied version of SQL [2] that does not require specication of table references and joins. The Schema-Free SQL system can expresses queries not supported in keyword-based searches [1]. Unlike Universal Relation approaches, the system is scaleable to large schemas, and it has built-in mechanisms for handling ambiguity and ranking interpreta-tions for the user. The auto-completion feature is not intended to remove all of the complexity in building SQL queries, just like auto-completion of code frag-ments does not remove the challenges of programming. However, it does make it easier to build SQL queries. Thus, the system provides a value-added feature to SQL querying that increases its exibility and usability with no sacrice in expressiveness or performance. The amount of the nal SQL code that is auto-completed depends on the number of joins and the complexity of the rest of the SQL expression. The time to complete a query takes around 5 milliseconds [3]. This demonstration also illustrates the value of identifying semantically equiv-alent join paths (shortcut joins) [3]. If not handled properly, a query with a single unique interpretation may have multiple equivalent SQL interpretations. Remov-ing this false ambiguity results in fewer queries being identied as ambiguous and reduces user confusion. In the demonstration we: { Show pre-entered queries on the TPC-H schema and their auto-completion. { Allow users to enter queries and have them completed by the system. { Demonstrate how removal of shortcut joins greatly reduces the amount of ambiguity and number of query interpretations. For TPC-H the number of unambiguous queries is increased by a factor of three [3]. Overall, the contribution is a very usable system for auto-completion of SQL, and a general technique (shortcut joins) for reducing query ambiguity that ca