Ingres 9.2 offers Unicode support
Version 9.2 of the free Ingres database enables users for the first time to process Unicode text. The database processes UTF-8 coded data for
long varchar data types. There is no conversion between the server and the client – both have to be set to UTF8. When creating tables, the lengths of Unicode columns have to be specified in bytes instead of characters.
In the new version of Ingres, developers can take advantage of the scrollable cursor. A scrollable cursor allows the server to process a results set in any order. The
OFFSET keywords have been added for
SELECT queries, which makes it possible to deliver partial returns of query results – for web applications, for instance. The statement:
SELECT given name, surname
ORDER BY surname
OFFSET 10 FETCH NEXT 10 ROWS ONLY
... returns data sets 10 to 20 in the
customer table sorted according to the
surname column. There are, however, some limitations when using the two new keywords: they may not be used in
VIEWs and are permissible in only a few situations in subselects.
There were also improvements to LOBs (large objects) and functions. Ingres’ JDBC and OpenAPI for the first time offer functions that allow direct access to LOB files – LOB locators, which can also retrieve part of a LOB. New mathematical functions include rounding operators,
floor. Strings can be trimmed using
rtrim, padded using
lpad, and modified using
A download of Ingres 9.2 is currently only available for Solaris; the previous version is also available for Linux, Windows, HP-UX and AIX. The source code is available from a subversion repository. Release notes and additional documentation are available in PDF format.