Project Metamorphosis: Wir präsentieren die Event-Streaming-Plattform der nächsten GenerationMehr Erfahren

Data Wrangling

Heterogeneous JSON Structures

Sometimes the messages on a Kafka topic will not have the same structure. For example:

  • A field is not present in each message
  • There are messages of multiple types in the topic, such as:
    { "Header": { "RecType": "RecA" }, "RAFld1": { "someFld": "some data", "someOtherField": 1.001 }, "RAFld2": { "aFld": "data", "anotherFld": 98.6 } }
    { "Header": { "RecType": "RecB" }, "RBFld1": { "randomFld": "random data", "randomOtherField": 1.001 } }

When defining a stream in KSQL against a topic with JSON data, there are some useful techniques to be aware of:

  • You can declare a schema for fields that are not present in every message. If a field is not present then a null is returned.
  • Use STRUCT for nested fields if you want to declare the schema for the contents too.
  • You can use VARCHAR against the parent element of a nested field, and then EXTRACTJSONFIELDFIELD function to access nested fields at execution time.
Visit the new Kafka-Tutorials site for the latest code examples


1. Inspect the raw data:

 ksql> PRINT 'source_data' FROM BEGINNING;
 {"ROWTIME":1545239521600,"ROWKEY":"null","Header":{"RecType":"RecA"},"RAFld1":{"someFld":"some data","someOtherField":1.001},"RAFld2":{"aFld":"data","anotherFld":98.6}}
 {"ROWTIME":1545239526600,"ROWKEY":"null","Header":{"RecType":"RecB"},"RBFld1":{"randomFld":"random data","randomOtherField":1.001}}

2. Register the source_data topic for use as a KSQL stream called my_stream:

 CREATE STREAM my_stream (Header VARCHAR, 
                          RAFld1 VARCHAR, 
                          RAFld2 VARCHAR, 
                          RBFld1 VARCHAR) 

3. Inspect the messages. Note that in the second message (which is record type “B”) there is no value for RAFld1 so a null is shown:

 ksql> SELECT Header, RAFld1 FROM my_stream LIMIT 2;
 {"RecType":"RecA"} | {"someOtherField":1.001,"someFld":"some data"}
 {"RecType":"RecB"} | null

4. Populate a new Kafka topic with just record type “A” values, using EXTRACTFROMJSON to filter record types on the Header value, and to extract named fields from the payload:

 SELECT EXTRACTJSONFIELD(RAFld1,'$.someOtherField') AS someOtherField, 
         EXTRACTJSONFIELD(RAFld1,'$.someFld')        AS someFld, 
         EXTRACTJSONFIELD(RAFld2,'$.aFld')           AS aFld, 
         EXTRACTJSONFIELD(RAFld2,'$.anotherFld')     AS anotherFld 
         FROM my_stream \
 WHERE EXTRACTJSONFIELD(Header,'$.RecType') = 'RecA';

Note that the serialization is being switched to Avro so that the schema is available automatically to any consumer, without having to manually declare it.

5. Observe that the new stream has a schema and is populated continually with messages as they arrive in the original source_data topic:

 ksql> DESCRIBE recA_data;

 Name                 : RECA_DATA
 Field          | Type
 ROWTIME        | BIGINT           (system)
 ROWKEY         | VARCHAR(STRING)  (system)
 For runtime statistics and query details run: DESCRIBE EXTENDED <Stream,Table>;

 ksql> SELECT * FROM recA_data;
 1545240188787 | null | 1.001 | some data | data | 98.6
< Back to the Stream Processing Cookbook