@agrosner that is a fair point, but is it worth the slowdown? The reason I bring this to public forum is I'm finding tuples to ironically increase readability and reduce boilerplate for a lot of grouping operations, which happens a lot in data science-y environments