aokolnychyi edited a comment on issue #173: Make collection of lower/upper bounds configurable URL: https://github.com/apache/incubator-iceberg/issues/173#issuecomment-498022864 I am wondering about the best way to implement this feature. A few ideas that came to my mind are below. **Option 1 (TableProperties)** We will have to add multiple new properties. For example, we might use something like this: ``` metrics.collection.enabled=true metrics.collection.column-delimiter=, metrics.collection.bounds.enabled=true metrics.collection.bounds.columns=col1,col2.nested ``` This option is a bit cumbersome and might become tricky (e.g., we need to reason about expressing nested columns, etc). Instead of storing a list of columns, we can store a regular expression. Then we will be able to construct a `MetricsSpec` based on those properties and pass it to writers. **Option 2 (Extend Metadata)** Alternatively, we can store a `MetricsSpec` directly in the metadata (similar to `PartitionSpec`) and expose an API to update it. Also, users will be able to provide a correct `MetricsSpec` while creating tables. The design of `MetricsSpec` is an open question. @rdblue any suggestions on this?
---------------------------------------------------------------- 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 --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
