[FEA] support rank and dense_rank for nested structs #8959
Labels
feature request
New feature or request
libcudf
Affects libcudf (C++/CUDA) code.
Spark
Functionality that helps Spark RAPIDS
Milestone
Is your feature request related to a problem? Please describe.
We are trying to push the Spark plugin to support as much of structs as possible. It would be nice to have rank and dense_rank support nested structs as input columns. This does not include lists as child columns under the structs.
Describe the solution you'd like
I don't know the best way to do this, but I suspect that the flatten method could be used with relatively small modifications to make this work.
Describe alternatives you've considered
Do the flatten ourselves, but it feels like it would be better for cudf to do it, especially if in the future we support other algorithms that do support lists or there are better, more optimized algorithms besides flatten.
Additional context
I know that @rwlee has been looking into this at least a little
The text was updated successfully, but these errors were encountered: