Comments
-
That's an option we'd explore as a last case scenario, though I'd be worried about runtime/job success when joining 150+150+100 column tables at 12M rows a piece. Short-term trying to figure out a column limit and going from there. May break these into use-case specific datasets from a single source anyway.
-
Tried 150 columns and it successfully completed after 32m 18s. Tried 200 columns and is failed after 47m 26s. Sounds like the sweet spot is somewhere in between.
-
Link to Support Case #05769383