Converting Julia DataFrame to R object using RCall.jl, one column at a time. String, Bool, Integer (32bit max in R), Real columns work properly. CAUTION: performance not tested for large objects.

julia_to_R_RCall.jl

## RCall data exchange between Julia and R
using DataFrames
using RCall

# example DataFrame
df = DataFrame(coln1 = 1:4, coln2 = ["M", "F", "F", "M"], coln3 = [false, true, false, true], coln4 = int8([1, 0, 1, 1]))

# R environment in a session started from Julia
g = globalEnv
reval(rparse("dfls <- NULL"))

# add columns one at a time converting Julia vectors to R-types via RCall.sexp
#  https://github.com/JuliaStats/RCall.jl/blob/master/src/sexp.jl
for cnm in DataFrames._names(df)
  g[:colcnm] = sexp(convert(Array, df[cnm]))
  reval(rparse("dfls\$$cnm <- colcnm"))
end
reval(rparse("df <- data.frame(dfls)"))
reval(rparse("save(file='dfjulia.RData', df)"))