Hi folks, I've run into https://github.com/delta-io/delta/issues/1499 where a VOID column gets created in the delta table, but can't be used from spark. (I did it by merging a dataframe with a new column containing only NULLs into an existing table). E.g. I'm trying to

merge

a dataframe into the table, and get error. For now I'm experimenting, and simply wiped the table and recreated. But were I to run into this issue in a situation where I could not lose data, I'm not sure what I should do.