hi!,
i have an ontology object which has a property derived from a column in a backing dataset, I want to use a new column to back that property but this new column does not have rows for the objects already created (it is an incremental build not a snapshot build); i need to do a one-off coalesce to backfill the rows of this new column based on the values of the old column, before i can migrate to the new backing column as this object property is being used in production - any ideas on how i can tackle this? (trying to stay away from approaches that will fundamentally create a different backing dataset, i.e. i don’t want to hit replace on the backing datasource in ontology manager)
You could up the semantic version of the incremental pipeline, and join this data into the new column, if that column is the source of all of the data and there aren’t any user edits.
If there are user edits, you could do a materialisation of the existing object type, and join the data from that property into the new (empty) column in the backing dataset.
Not sure how that will integrate with your existing pipelines, but it should be doable.