cudf: min()/max() on empty series returns a large/small value instead of nan

Reporting a bug

• I am using the latest version of PyGDF from conda or built from master.
• I have included the following environment details: Linux Distro: Ubuntu 16.04, Linux Kernel: Linux 4.15.0-29-generic x86_64 , GPU Model: TITAN V
• I have included the following version information for: Arrow: 0.7.1, CUDA: 9.2.148, Numpy: 1.14.5, Pandas: 0.23.3, Python: 3.6.6
• I have included below a minimal working reproducer (if you are unsure how to write one see http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports).

Calling min() or max() on an empty series returns largest/smallest value for the dtype of the series respectively.

import pygdf
from pygdf import Series

gseries = Series([])
print(gseries.min())
print(gseries.max())

Output:

1.7976931348623157e+308
-1.7976931348623157e+308

Pandas returns nan for similar cases.

---

Workaround: Check if the series is empty before calling min()/max()