The following provides usage information for the Confluent SMT io.confluent.connect.transforms.ExtractTopic.


Extract data from a message and use it as the topic name. You can either use the entire key/value (which should be a string), or use a field from a map or struct. Use the concrete transformation type designed for the record key (io.confluent.connect.transforms.ExtractTopic$Key) or value (io.confluent.connect.transforms.ExtractTopic$Value). You can also extract the entire value from a message header value (string) by using the concrete type (io.confluent.connect.transforms.ExtractTopic$Header).


This transformation is developed by Confluent and does not ship by default with Apache Kafka® or Confluent Platform. You can install this transformation via the Confluent Hub Client:

confluent-hub install confluentinc/connect-transforms:latest


The configuration snippet below shows how to use and configure the ExtractTopic SMT.

transforms=KeyExample, ValueFieldExample, KeyFieldExample, FieldJsonPathExample, HeaderExample

Use the key of the message as the topic name.


Extract a required field named f2 from the value, and use it as the topic name.


Extract a field named f3 from the key, and use it as the topic name. If the field is null or missing, leave the topic name as-is.


Extract the value of a field named f3 in the f1 field in the key, and use it as the topic name. Here the format of the field is defined with JSON Path (e.g., ["f1"]["f3"]). If the field is null or missing, leave the topic name as-is.


Extract the value of a message header (as a string) with key h1 (required) and use it as the topic name.



Name Description Type Default Valid Values Importance
field Field name to use as the topic name. If left blank, the entire key or value is used (and assumed to be a string). string “”   medium
field.format Specify field path format. Currently two formats are supported: JSON_PATH and PLAIN. If set to JSON_PATH, the transformer will interpret the field with JSON path intepreter, which supports nested field extraction. If left blank or set to PLAIN, the transformer will evaluate the field config as a non-nested field name. When using ExtractTopic$Header, only the default PLAIN format can be used, which will extract the header value as a string. string “PLAIN” “JSON_PATH”, “PLAIN” medium
skip.missing.or.null How to handle missing fields and null fields, keys, and values. By default, this transformation will throw an exception if a field specified via the field configuration is missing or null, or if no field is specified but the message’s key or value is null. If this configuration is set to true, the transformation will instead silently ignore these conditions and allow the record to pass through unaltered. boolean false   low


Transformations can be configured with predicates so that the transformation is applied only to records which satisfy a condition. You can use predicates in a transformation chain and, when combined with the Filter (Apache Kafka), predicates can conditionally filter out specific records. For details and examples, see Predicates.