For data types like INT, LONG etc, there is a default value given in the doc. We have a use case where we have columns of these data types being null. Is it somehow possible to keep them null instead of INTEGER.MIN_VALUE and so on? We are ingesting data from kafka topic into a RT table.

Have a look at https://docs.pinot.apache.org/developers/advanced/null-value-support

Thanks, but we had been facing issues with nullHandlingEnabled and table version while querying on presto. The error went as:

ERROR [PinotClientRequest] [jersey-server-managed-async-executor-3] Caught exception while processing POST request
java.lang.IllegalStateException: Null handling cannot be enabled for data table version smaller than 4.

I had configured all the multi-stage engine configs everywhere in cluster, server and broker configs but the error persisted. Even I couldn't run joins on pinot as on using v2 engine it would give an error: "Error composing query plan". I had seen a thread with similar issue where it was deduced that release v11 has issues with multi-stage engine. Could you please confirm if this is the issue and suggest anything? here is the thread: https://apache-pinot.slack.com/archives/C03Q4A11GC9/p1665428747644939