This is a known issue in SilkPerformer 6.0. In special circumstances the ODBC Recorder does not detect the correct number of fetched rows. The incorrect rows of data are printed into the Fetch Data Table of the script causing a memory reference error. Situations when this happens:- If the attribute SQL_ROWSET_SIZE is used to define the rowset count and then SQLFetchScroll or SQLFetch using an ODBC 3 driver is called. If the attribute SQL_ATTR_ROW_ARRAY_SIZE is used to define the rowset count and then SQLExtendedFetch is called . The above described behaviour occurs only if no Row Fetches Pointer is set to retrieve the number of fetched rows. Information on how ODBC handles the fetching of data:- SQLFetch normally fetches one single row. SQLExtendedFetch fetches a number of rows per call. The number of rows is defined by the rowset count. The rowset count is set with the attribute SQL_ROWSET_SIZE. SQLFetchScroll fetches a number of rows per call. The number of rows is defined by the rowset count. The rowset count is set with the attribute SQL_ATTR_ROW_ARRAY_SIZE. The SQLFetch call is mapped to SQLExtendedFetch in ODBC 2, and therefore fetches a number of rows defined by the attribute SQL_ROWSET_SIZE. If the driver does not support SQLExtendedFetch, then a single row is fetched. The SQLFetch call fetches a number of rows defined by the attribute SQL_ATTR_ROW_ARRAY_SIZE in ODBC 3. If this problem is encountered, please contact Support for a patch for SilkPerformer 6.0. This will be fixed in the upcoming Service Pack.
↧