Schema-Free SQL- Providing Freedom from The Schema Knowledge Required to Formulate SQL Queries

The current standard query language SQL requires users to know how data isnormalized and stored in a database schema to compose queries. One of themajor challenges in generating these queries is the identification of joins toconnect attributes located in different relations. Constructing these joins inSQL or through a graphical query interface is tedious error-prone and notintuitive to casual users. Any new query language should maintain the expressivenessof SQL without requiring schema knowledge or an additionalmodel. A new Schema-Free query system is developed to support theautomatic completion of an SQL query without the full specification of joinsand tables. A new algorithm efficiently constructs all maximal sets of losslessjoins in a schema an order of magnitude faster than previous methods. Thequery system uses a new query language allowing under-specification of SQLqueries to remove the burden of schema knowledge from the user whilecontinuing to support the full range of SQL queries. Queries with multipleinterpretations are disambiguated through three different approaches.