bowenli86 commented on issue #8831: [FLINK-12935][hive] package flink-connector-hive into flink distribution for HiveCatalog URL: https://github.com/apache/flink/pull/8831#issuecomment-505523726 Hi @zentol , we decided a long time ago with Fabian and Timo to have HiveCatalog in flink-connector-hive because it's a metadata connector. And all components (metadata connector, source, sink, hive udf) in flink-connector-hive shares lots of code. Moving it to its own module just because of the concern of Flink distribution size seems a bit overkill to me. One critical functionality of HiveCatalog is to enable users to persist their Flink metadata (table, view, partition, udf, etc) in Hive metastore, and I think including it in Flink distribution is important to the user experience completeness for Table API and SQL. flink-connector-hive jar is only 117kb. Yeah, I also resonate with you on the concern that other connectors may try to follow and squeeze in. How about adding a `/catalog` dir within `/opt` to bring up attention that adding flink-connector-hive is not about data connector but catalog/metadata connecter? Not many other connectors would fit in this requirement.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
