Pandas read_csv with chunksize stops before end of file

[jcrist]

I have a massive csv file (54 GB, ~2.7 billion records) that I'm trying to process in chunks using chunksize. The csv reader stopped before hitting the end of the file. Upon further investigation, I found that it stopped after reading 2147483648 rows, which, is (2**32)/2. I assume there's an overflow happening somewhere inside the cython code for read_csv.