john-bodley commented on issue #14909: URL: https://github.com/apache/superset/issues/14909#issuecomment-854296634
@betodealmeida thanks for the excellent worded SIP. I'm in agreement with the direction. A few comments: **Tables** 1. Regarding tables when/how are these added? 2. Is there any plan for these to be regularly synced to ensure they accurately represent the underlying data warehouse (see [SIP-7](https://github.com/apache/superset/issues/5842) for specifics)? 3. If (2) How are differences resolved between columns in `Table` and `Dataset` models? **Datasets** 1. The `source_type`/`source_id` polymorphic association is somewhat akin to the current `datasource_type`/`datasource_id` association in the `Datasource` model, used primarily to differentiate between a Druid NoSQL datasource and SQL table where the `datasource_type` determined which table one should join with. Having this necessary explicit (weak) relationship rather than an implicit (strong) relationship (courtesy of a foreign-key) has never seemed desirable from a data integrity perspective, and with the deprecation of Druid NoSQL I was hoping we could remove said pattern. I realize there is no was to define a polymorphic associations using SQL constraints given foreign key constraint always references one target table, but I was wondering if one of the alternative options outlined [here](https://stackoverflow.com/a/441111) would be more desirable, i.e., a super-table or using two columns (preferred for simplicity reasons). -- 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]
