Hi,
Here are the questions I think we should consider:
1. Do we need this at all given that we have the partition argument in
ProducerRecord which gives full control? I think we do need it because this
is a way to plug in a different partitioning strategy at run time and do it
in a fairly
Hey Harsha,
A few comments:
Can you finish up the KIP there are some unfinished sentences and odd
whitespace things going on.
Here are the questions I think we should consider:
1. Do we need this at all given that we have the partition argument in
ProducerRecord which gives full control? I
Hi Jay,
Sorry about the KIP formatting . I fixed those in the KIP.
2. We certainly need to add both the serialized and unserialized form for
the key as both are useful.
I added those to the interface.
3. Do we need to add the value? I suspect people will have uses for
computing
Hey Sriharsha,
Great, thanks!
For 4:
Yeah the use case for init and close is making use of any kind of metadata.
An example of this would be if you are trying to do range partitioning you
need to map lexicographic ranges to numeric partitions. You might do this
by adding a new property to the
Hi,
Here is the KIP for adding a partitioner interface for producer.
https://cwiki.apache.org/confluence/display/KAFKA/KIP-+22+-+Expose+a+Partitioner+interface+in+the+new+producer
There is one open question about how interface should look like. Please take a
look and let me know if you
