If you need to group-by an attribute/field in the value, you will need to extract the grouping attribute/field and set it as message key. Afterwards, you write it back into a topic to repartition the data accordingly and read it back to process it.
-Matthias On 7/18/19 2:41 PM, Navneeth Krishnan wrote: > Hi All, > > I'm working on a kafka streams application to basically aggreagte data over > 5 min intervals. > > The input topic is partitioned by id and then I want to use process > function to aggregate data using state store and use punctuator to emit > results. But I'm not sure how I can perform groupBy when using low level > api's. My intention is to achieve the below topology. > > Topology: Source -> Map values -> GroupBy (internal key) -> Process -> Sink > > Need some help on how I can achieve this. Appreciate all the help. > > Thanks >
signature.asc
Description: OpenPGP digital signature
