> it would be invalid to specify a custom extractor (for correctness reasons, we enforce the default timestamp extractor)
"default timestamp extractor" meaning the one that was used when reading the input record? So, KStream.through() and KStream.to()-Topology.stream() are not semantically equivalent? Is there anything wrong with using the latter in order to specify a timestamp extractor? My motivation is to "fix" timestamps that result from out-of-order operations until https://issues.apache.org/jira/browse/KAFKA-6454 arrives. On Fri, Feb 2, 2018 at 7:22 PM, Matthias J. Sax <matth...@confluent.io> wrote: > It's not required to specify Consumed: > > - the specified Serdes from Produced are used for the consumer, too > - it would be invalid to specify a custom extractor (for correctness > reasons, we enforce the default timestamp extractor) > > Note, that Streams API sets the metadata record timestamp on write, and > thus, using the default timestamp extractor ensures that this timestamp > is correctly reused when reading data back. > > > > -Matthias > > On 2/2/18 3:54 PM, Dmitry Minkovsky wrote: > > `KStream#through()` currently lets you specify a `Produced`. Shouldn't it > > also let you specify a `Consumed`. This would let you specify a time > stamp > > extractor. > > > >
