The implementation of the RWDBCursor class in the DBTools.h++ access library for ODBC is subject to many restrictions. In most cases, these restrictions are imposed by the capabilities of the underlying driver.
Some drivers require a unique index on every table involved in every cursor operation; e.g., Microsoft / Sybase SQL server. An error from the ODBC driver will result if any table lacks a unique index when required.
Scrolling cursors are available through the ODBC access library, and are implemented in the access library by setting the SQL_ATTR_CURSOR_SCROLLABLE statement attribute to SQL_SCROLLABLE. This is done automatically when a cursor of type RWDBCursor::Scrolling is requested.
Scrolling cursors and sequential cursors act differently in respect to large blob and string data. Scrolling cursors are not able to fetch large blob or string data of unlimited size. For scrolling cursors, memory is allocated and bound to receive the column data. Data will be fetched only up to the size of the allocated memory. There are RWDBODBCLibEnvironmentHandle member functions that control the size of these allocations. Sequential cursors are not restricted in this way.
©Copyright 2000, Rogue Wave Software, Inc.
Contact Rogue Wave about documentation or support issues.