>>> %timeit df2 = table.to_pandas(nthreads=1)
10 loops, best of 3: 158 ms per loop