Colin Hicks

@stebl, thanks for filing this issue. I'm also not aware of an issue tracking this limitation, which also applies to other join types. Checking my understanding of the scenario and...

I believe the cause of this behavior is that the generated Protobuf schema uses the corresponding scalar type, rather than the wrapped, nullable Protobuf type. For example `name VARCHAR` is...

Update: I had overlooked that we introduced a knob to control the wrapping behavior in https://github.com/confluentinc/ksql/pull/8934, with the internal `unwrapPrimitives` property. Primitives are used for new queries after ksqlDB 0.24....

I elevated the priority based on product feedback.

This was reported by a user in the Slack community. They reported the change was breaking their application after upgrading from an older Confluent Platform version. It's also a best...

Yeah, I understand the potential confusion with `UNIX_TIMESTAMP`. This function was introduced before KSQL had a `timestamp` type. When timestamp and date types were introduced, we also introduced utility functions...

> 2. In the case that the topic exists and retention is different (e.g. in the CS case), do we want to consider either (a) failing the statement or (b)...

There is an existing ticket to unify these configs so that there is less confusion moving forward: https://github.com/confluentinc/ksql/issues/7351. I think we should not change any defaults until we tackle that.

I think another source of the confusion is that Confluent Cloud KSQL does in fact configure `snappy` (for both configs here), while other KSQL distributions do not specify the compression.

Closing this. As noted above, the upstream change was [reverted](https://github.com/apache/kafka/pull/12383), pending a reimplementation of KIP-770. We can reopen this at a later point.