The PutMarkLogicRecord
processor differs from the PutMarkLogic
processor in that instead of inserting a document per FlowFile, it breaks a FlowFile into batches of documents that are inserted into MarkLogic. The processor depends on a NiFi RecordReader
for controlling how multiple records are extracted from a single FlowFile. The processor also depends on a NiFi RecordWriter
for controlling how records are serialized before they are written as documents to MarkLogic.
The processor requires a MarkLogic user with the rest-writer
privilege.
The primary use case for this processor is when an incoming FlowFile should result in multiple documents being written to MarkLogic as opposed to just one document being written. A FlowFile representing a CSV file is a common example of when this processor would be used.
An example of this processor in action can be found in the Getting Started guide.
For full details on each property and relationship for this processor, add an instance of it to the NiFi canvas, right-click on it, and select “View Usage”.
Restart support
Please see the PutMarkLogic
documentation for information on the restart support added in the 1.24.2 release of the connector.