>>> %timeit df2 = table.to_pandas(nthreads=4)
10 loops, best of 3: 103 ms per loop