DataFrames.jl
DataFrames.jl copied to clipboard
investigate using Int32 in joins
If both tables in join-operations are fitting into Int32
range (probably a common case) then we can use Int32
instead of Int
for indexing. This should speed up things.