c-thiel commented on issue #591: URL: https://github.com/apache/iceberg-rust/issues/591#issuecomment-2474101258
This is not so much about this specific use case - which I also don't care about much either, but about having two different representations for the same entity. Took me a while to formulate it so clearly. There are many lines of code in different files just to work around this problem. Let's assume for example we want to add a field to a schema, then the point representation in the `PartitionSpec` is not compatible with the one in the `Schema`. Java would just throw an exception in this case. Because we have two distinct representations for the same entity, and those representations are not bijective, we need extra code to handle conversion. https://github.com/apache/iceberg/blob/e06b069529be3d3d389b156646e751de3753feb0/core/src/main/java/org/apache/iceberg/SchemaUpdate.java#L97-L103 also https://github.com/apache/iceberg/blob/e06b069529be3d3d389b156646e751de3753feb0/core/src/main/java/org/apache/iceberg/SchemaUpdate.java#L112 several lines of docs here: https://github.com/apache/iceberg/blob/e06b069529be3d3d389b156646e751de3753feb0/api/src/main/java/org/apache/iceberg/UpdateSchema.java#L54-L58 We are even inclined to document at [some places](https://github.com/apache/iceberg/blob/e06b069529be3d3d389b156646e751de3753feb0/api/src/main/java/org/apache/iceberg/UpdateSchema.java#L95) that a point is nothing special here: ``` * <p>The given name is used to name the new column and names containing "." are not handled * differently. ``` All of this could be solved by using a globaly unique identifier for a field - which for me is very clearly `Vec<String>` and just handle conversion to a string only representation at the last moment. I stumbled across this again when implementing `SchemaUpdate`. So lets forget about the use-case, but maybe still consider harmonizing identifiers for a field. -- 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. To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org For additional commands, e-mail: issues-h...@iceberg.apache.org
