Record Reader property identifies the controller service that the processor uses to deserialize incoming data into NiFi records. You select an appropriate reader based on the format of the input data.
(This older Record-Oriented Data with NiFi blog describes the Apache NiFi processors and controller services available for working with record-oriented data.)
You configure a
Record Reader via the processor configuration dialog, PROPERTIES tab. You can also add a new reader instance via the Operate Palette configuration dialog CONTROLLER SERVICES tab.
The Connector supports all compatible
Record Reader controller services, and has been specifically tested with certain data formats. These readers, data formats, and their schema origins are identified in the table below.
|Reader Name||Data Format||Schema||Description|
|AvroReader||Avro||Embedded in the Avro data, obtained from a schema registry, or explicitly specified.||Each Avro record is deserialized to a NiFi record.|
|CSVReader||CSV||Inferred from the data, obtained from a schema registry, or explicitly specified.||Each row is deserialized to a NiFi record.|
|JsonTreeReader||JSON||Inferred from the data, obtained from a schema registry, or explicitly specified.||Each JSON record is deserialized to a NiFi record.|
|ParquetReader||Parquet||Embedded in the Parquet data.||Each Parquet record is deserialized to a NiFi record.|
|XMLReader||XML||Inferred from the data, obtained from a schema registry, or explicitly specified.||The second level (within enclosing