Physical Database Characteristics

---

Are Multiple character sets allowed? The SQL3 standard [29] specifies a syntax to enable any column to be encoded in any character set with any default collation. Jim Melton, one of the creators of this part of the standard states that it was a necessary evil until Unicode became prevalent and a single canonical form could be used [28]. ADABAS and Interbase provide a syntax and possible encodings that come closest to SQL3.

The figure above shows different places where a character set encoding can be used for storage according to SQL3. The primary domain or database, the table, and the column. What the author found is that virtually every database supplier interviewed will offer the ability to keep the default data for the database and tables in one character set (which may be any character set, including UTF-8) while enabling storage of UCS-2 data alongside it with an alternate datatype of some sort. The exact implementation varies from vendor to vendor.