Github user tomatophantastico commented on the issue:
https://github.com/apache/metamodel/pull/178
wrt 1) no, not required. if you got a use case, then this is fine with me.
I just think, for getting the data from the value, you still need a lot of
domain knowledge. I just wondered if there is a declarative way to make the
value accessible.
wrt 2) i disagree that this holds for cassandra, as cassandra features a
rich query api (and cql) and beyond that a some sort of schema. in Kafka the
message carries in most cases a binary blob and you will need to iterate over
all of them.
It is true that for all postprocessingdatacontexts (virtually all but JDBC)
join and other operations are resource intensive, as they are performed in
memory.
The method you propose is fine for the infrequent case, small data set and
simple lookups. Metamodel otoh features a rich query api, so performance issues
can surface quickly.
As this is the case with other systems, i suppose this is ok. Kafka just
struck as it is a mix of remote data locality, potentially huge size and lack
of query capabilities.
---
