dawidwys commented on issue #9203: [FLINK-13375][table-api] Move ExecutionConfigOptions and OptimizerConfigOptions to table-api URL: https://github.com/apache/flink/pull/9203#issuecomment-514185439 Generally speaking I also prefer the `table` prefix. I agree with @twalthr that in Flink's naming conventions the module that works with the relational algebra was usually referred as a "table ecosystem" or "Table API", of which the SQL language is part. Some additional comments regarding the @wuchong's points Ad.1 I am not aware of any decisions that we want to make the SQL string representation as the go-to interface. I think what we want is for the relational Table API (both sql as well as the java/scala counterparts) to be the first class citizen among different APIs like DataStream/DataSet. Ad.2 To be honest I could revert the point with just substituting `table.` with `sql.` ;) Does it mean the sql in `.sqlUpdate`? Does it mean the `Table API` or just the sql subpart? Ad.3 I don't think this is a problem. Moreover I think some of the mentioned options actually abuse the term `table`. E.g. in `sql.resource.hash-agg.table.memory.mb` I think it actually refers to managed memory of the hash aggregate operator. What does the `table` there means? On the other hand in the option `sql.optimizer.reuse.table-source.enabled` there is no table, only `table-source`. Theoretically we could even think about dropping the `table-` from `table-source` in that option.
---------------------------------------------------------------- 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
