[
https://issues.apache.org/jira/browse/KAFKA-3729?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15379721#comment-15379721
]
Guozhang Wang commented on KAFKA-3729:
--------------------------------------
Currently there are two ways to specify serde classes in Kafka Streams: 1) you
can specify a "default" serde through the StreamsConfig as "key.serde",
"value.serde", etc; 2) you can override the serde in the Streams DSL on a
per-operator granularity as well, for stateful operators like join and
aggregate. If you use the overloading functions without the overridden serde
class, or specify as null, the default ones in the config are still used.
Many serializer / deserializers need to be configured before use (think:
AvroSerializer), and the Kafka Serde interface provide such API for users to
specify the configuration logic. This "configure" API will be automatically
triggered for case 1) above, but not for case 2), hence users need to do sth.
like
{code}
Serializer mySer = new MySerClass();
mySer.configure(properties);
KStream joined = stream1.join(stream2, ...., mySer);
{code}
which is not convenient; we can let the library to auto-trigger the configure
function for case 2) as well so that the above example can get ride of the user
manual triggering configure function line above.
Bharat, does this make sense to you?
> Auto-configure non-default SerDes passed alongside the topology builder
> ------------------------------------------------------------------------
>
> Key: KAFKA-3729
> URL: https://issues.apache.org/jira/browse/KAFKA-3729
> Project: Kafka
> Issue Type: Improvement
> Components: streams
> Reporter: Fred Patton
> Assignee: Bharat Viswanadham
> Labels: api, newbie
>
> From Guozhang Wang:
> "Only default serdes provided through configs are auto-configured today. But
> we could auto-configure other serdes passed alongside the topology builder as
> well."
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
