SQL92 - Internationalization Issues and Proposed Solutions

The ever popular Structured Query Language (SQL) has continued to be the standard language to access relational databases for almost two decades. As SQL is widely used by all major relational database management systems (RDBMS), there is an ongoing effort to standardize syntax and semantics in the SQL (ISO/IEC) standards SQL-92 and SQL3. The first part of this presentation will present and discuss current language support concepts and features in the ANSI SQL standards including character sets and encodings in SQL data types, identifiers and literals, translations, collations, conversions, string functions, error handling and messaging, and call-level interfaces. It will be shown that those features, as specified in the standards, are still not mature and need to be refined before being used by actual implementations. The second part of the presentation proposes changes and additions to make language support features in the SQL standards workable. Topics covered will include a discussion of using Unicode as the basic underlying character set for SQL, addition of a standard Unicode datatype, language tagging, Unicode-based collation concepts, handling text-elements versus individual characters, use of standardized character set and collation names, Unicode-based string functions, and other topics. In the writing of this paper, two major database vendors have worked together with the original authors of the international features portions of the SQL Standard to come up with a proposal that meets user's needs, is feasible to implement, and poises SQL for expanded portable use in a global environment.