When you process an XML file or stream, the XML parser parses the entire XML file and writes target XML data at end of file. Use the On Commit attribute to periodically flush the data to the target before reaching end of file. You can flush data periodically into one target XML document, or you can generate multiple XML documents.
You might want to flush XML data in the following situations:
Large XML files.
If you are processing a large XML file of several gigabytes, the Integration Service may have reduced performance. You can set the On Commit attribute to Append to Doc. This flushes XML data periodically to the target document.
. If you process real-time data that requires commits at specific times, use Append to Doc.
You can set the On Commit attribute to one of the following values:
Generate and write to the XML document at end of file.
Append to document.
Write to the same XML document at the end of each commit. The XML document closes at end of file. This option is not available for XML Generator transformations.
Create new document.
Create and write to a new document at each commit. You create multiple XML documents.
You can flush data if all groups in the XML target are connected to the same single commit or transaction point. The transformation at the commit point generates denormalized output. The denormalized output contains repeating primary key values for all but the lowest level node in the XML schema. The Integration Service extracts rows from this output for each group in the XML target.
You must have only one child group for the root group in the XML target.