In the DBTools.h++ paradigm, all results are tables. Class RWDBTable encapsulates a single table of results, class RWDBResult encapsulates a sequence of 0 or more tables, and class RWDBReader provides the mechanism for reading a table. Consequently, handling multiple result sets is never a problem.
Since Adaptive Server supports the notion of a trigger, it is possible for the INSERT, UPDATE, and DELETE operations to produce result sets, as well as the more common SELECT. In DBTools.h++, the execute() method of each class returns an RWDBResult object. Applications that are concerned with results being returned by an INSERT, for example, can do so in a portable way by checking the RWDBResult returned by RWDBInserter::execute().
Data is exchanged between Adaptive Server and DBTools.h++ via the Client-Library calls ct_command(), ct_send(), ct_results(), and ct_fetch(). Before calling ct_command(), DBTools.h++ calls ct_cancel(CS_CANCEL_ALL); before processing the next result set, DBTools.h++ calls the function ct_cancel(CS_CANCEL_CURRENT). The net effect is that any data not processed by an application is silently discarded; DBTools.h++ applications do not need to be concerned with the state of the database connection.
©Copyright 2000, Rogue Wave Software, Inc.
Contact Rogue Wave about documentation or support issues.