rymurr commented on pull request #1843: URL: https://github.com/apache/iceberg/pull/1843#issuecomment-737233889
Just a bit of clarification on `PathIdentifier` and your above comment @rdblue: You are suggesting that the change in #1783 returns a `PathIdentifier` from the `SupportsCatalogOptions.extractIdentifier` method? I like that and it simplifies some of the logic in this PR. If I understand correctly in the case of stored procedures or other places where we have control we can use `PathIdentifier` or use the other mechanisms available to pass path information. Correct? What I am unsure about is how to handle path based `CREATE`, `DROP`, `ALTER` comamands. We don't have much control over the `Identifier` and we would still have to inspect if its a path. Unless we do a `CREAT TABLE x AS iceberg LOCATION path`, which works but the table name is redundant/ignored. Or are you suggesting this PR doesn't handle path based DDL coming from Spark? FWIW I agree that we shouldn't have a 'special' namespace for importing path based tables, however my approach of ignoring namespaces if the name is a path probably isn't quite right either. ---------------------------------------------------------------- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
