Hi Jeyhun, You are right. As long as the types in Tuple-n are known types(e.g., primitives) and the serdes for them are available you can remove the need for providing serdes and infer them. Indeed we have a project that we use similar approach to eliminate the need to have user to provide the serdes.
Thanks. --Hojjat On Tue, Oct 4, 2016 at 4:07 PM, Guozhang Wang <wangg...@gmail.com> wrote: > > ---------- Forwarded message ---------- > From: Jeyhun Karimov <je.kari...@gmail.com> > Date: Mon, Sep 19, 2016 at 1:15 AM > Subject: Streams support for Serdes > To: "dev@kafka.apache.org" <dev@kafka.apache.org> > > > Hi community, > > When using kafka-streams with POJO data types we write our own > de/serializers. However I think if we have built-in Serdes support for > Tuple-n<Types...> data types (ex:Serdes.Tuple2<Long,Double>) we may easily > use Tuples and built-in Serdes can help to reduce the development cycle. > Please correct me if I am wrong or if there is similar solution within > library please let me know. > > Cheers > Jeyhun > -- > -Cheers > > Jeyhun > > > > -- > -- Guozhang >
