Unable to read parquet file with fastparquet but works with pyarrow - nullable ints

Viewed 782

Currently running some code like this:

df = pd.read_parquet('/tmp/my-file.parquet', engine='pyarrow')

I was having memory consumption issues since the files are large so I wanted to investigate is fastparquet would work better for memory usage.

When I switch the engine:

df = pd.read_parquet('/tmp/my-file.parquet', engine='fastparquet')

This line now throws an error like this:

TypeError: int() argument must be a string, a bytes-like object or a number, not 'NoneType'

I believe this is happening because I have an integer field that contains nulls. I can't find any documentation that says this is not supported.

Any ideas on why this is happening or how to workaround while still using fastparquet?

0 Answers
Related