villebro commented on PR #36368: URL: https://github.com/apache/superset/pull/36368#issuecomment-3849971147
> I have a small nit with this approach. We are basically saying we won't need to filter tasks by these properties right? It looks like we could go with the columns approach with only downside being migration requirement ( which i think is reasonable .) Thanks for reviewing the proposal @msyavuz! I'm not super opinionated on this, so if there's pushback against the blob-first approach I'm open to moving these into dedicated fields. While working on this I just noticed there will likely be a need for adding new fields during migration to GTF, and I didn't want to burden those migrations with having to do a db migration every time a new property is needed. And adding them up front would likely have missed a few, requiring migrations anyway. As a compromise how does this sound: we go with blob approach for now, and once we've migrated all legacy tasks to GTF and have a full understanding of fields required for internal use cases, we do a migration that breaks them out? -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
