Implementing User-Defined Output Control
To implement user-defined output control within a user-defined Datapump application, you must define a class that implements the
rpsDp_UpdateCallback(tranID: Integer64; timestamp: TimeStamp; oid: Object; operation: Integer; tableName: String; dropRowOperation: Boolean io; outputRowInfo: JadeDynamicObject io) updating; vars count : Integer; indx : Integer; int : Integer; jdoPropName : String; jdoPropValue : String; begin // only Company is set as OutputControl in the RPS mapping, // so only rows for the table Company will be passed to this routine count := outputRowInfo.propertyCount(); // log information about updates to this table writeToMyLog("update callback=" & tranID.String & "; operation=" & operation.String & ";timestamp=" & timestamp.String & ";oid=" & oid.String & ";tableName=" & tableName); writeToMyLog('---' & outputRowInfo.getName & ' ' & outputRowInfo.getOidString & " " & count.String & ' ---'); if operation = JadeRpsDataPumpIF.ObjectDeleted then return; endif; // log information about each column in row foreach indx in 1 to count do jdoPropName := outputRowInfo.getPropertyName(indx); jdoPropValue := outputRowInfo.getPropertyValueByIndex(indx).String; log.info(jdoPropName & " = " & jdoPropValue & "."); // look for column for "phone" and update value if jdoPropName = "phone" then if jdoPropValue.String = "(64) 3 555 4519" then outputRowInfo.setPropertyValueByIndex(indx, "(64) 3 999 4519"); writeToMyLog(" phone changed to (64) 3 999 4519"); endif; endif; endforeach; // if required, set dropRowOperation parameter to true to not output the // row operation to the relational database end;
The
For create and update operations, the attribute values are populated with the values obtained from the JADE object, including values returned by user-defined column-mapping methods.
The application callback can interrogate attributes and change attribute values within the
Your data pump application must first call the
When database tracking commences, your user-defined Datapump application is driven by callbacks invoked in a defined transaction replication cycle.
Not all tables are likely to require output manipulation. To avoid the overhead of creating the dynamic objects and calling the user routine when not required, the Relational Population Service wizard enables you to select the tables for which the callback will be used.
By default, tables are not included in callbacks. (For details, see "
Selecting output callback for a table has no effect if callbacks are not registered in your Datapump application or if you have not defined your own Datapump application.