Cudf: [BUG] cudf.read_orc reads incorrect data for one row

[trstovall]

Describe the bug
A clear and concise description of what the bug is.

Steps/Code to reproduce bug

>>> df = cudf.read_orc('to_orc_bug.orc')
>>> df.upc_nbr[(df.visit_nbr == 14600028) & (df.store_nbr == 47)] = 681131184420
>>> df.to_orc('to_orc_bug.orc', compression='snappy')
>>> df2 = cudf.read_orc('to_orc_bug.orc')
>>> df2.upc_nbr[(df2.visit_nbr == 14600028) & (df2.store_nbr == 47)]
999786    2526351652

Expected behavior
Returned value should be 681131184420, not 2526351652.

Environment overview (please complete the following information)

• Method of cuDF install: Conda
• 0.14 nightly on ~ May 29

to_orc_bug.orc.zip

[kkraus14]

cc @devavret in case #5324 is related

[devavret]

Tried to read with pyarrow and it works.

import cudf
import pyarrow.orc as orc

df = cudf.read_orc("to_orc_bug.orc")
df.upc_nbr[(df.visit_nbr == 14600028) & (df.store_nbr == 47)] = 681131184420
df.to_orc("to_orc_bug2.orc", compression="snappy")

pdf = orc.ORCFile("to_orc_bug2.orc").read().to_pandas()
print(pdf[(pdf.visit_nbr == 14600028) & (pdf.store_nbr == 47)])

999786    6.811312e+11
Name: upc_nbr, dtype: float64

Seems to be a reader issue.

[kkraus14]

Relabeled issue as such

[devavret]

This was an easy fix but I'm still trying to figure out how to properly add tests for this, or in general, anything in cuIO.