Please see Database-like ops benchmark
There is a new Join benchmark where Julia’s DataFrames.jl is in the last 2 spots most of the time. And takes is 30x slower than R’s {data.table} in some of the benchmarks.
This reflects the fact that Julia’s data ecosystem is not as mature and Julia’s grouping benchmark was in a similar spot not long ago but has since improved greatly.
Hope to see more development in Julia’s data ecosystem. I want to see if Julia can match data.table in some cases. I have done some work on group-bys and managed to match data.table’s performance in some cases.