This rewrites some of the Later and DplyFrame code to greatly speed up mutates on grouped dataframes. In some cases I am seeing a roughly 10x improvement (5000 groups on the 54k row diamonds dataframe). I think we can further improve this performance, hopefully by another 10x, in the future by being smarter with the code I've added here.
This rewrites some of the
LaterandDplyFramecode to greatly speed up mutates on grouped dataframes. In some cases I am seeing a roughly 10x improvement (5000 groups on the 54k row diamonds dataframe). I think we can further improve this performance, hopefully by another 10x, in the future by being smarter with the code I've added here.