All parsers retrieve column data from the message text and write it to the Oracle GoldenGate trail. In some cases the columns are read in index order as defined by the source definitions, in other cases they are accessed by name.

Depending on the configuration and original message text, both before and after or only after images of the column data may be available. For updates, the data for non-updated columns may or may not be available.

All column data is retrieved as text. It is converted internally into the correct data type for that column based on the source definitions. Any conversion problem will result in an error and the process will abend.