[Python] CSV ReadOptions column_names None defaults to an empty array

[alikleit]

Describe the bug, including details regarding any error messages, version, and platform.

When using the ReadOptions type, and setting None to column_names. ReadOptions is then setting it to [].

Changing the value afterwards results into an error:

read_options.column_names = None
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

>   ???
E   TypeError: 'NoneType' object is not iterable

This means that it force us to define our columns manually, rather than it automatically discovering it from the head.

Any idea?

Component(s)

Python

[raulcd]

Thanks for raising the issue! Is this about pyarrow.csv.ReadOptions? Are you using autogenerate_column_names?
Could you share a small snippet of your use case so it's easier for us to understand what is the use case you are looking for?
I basically don't understand what you mean by:

This means that it force us to define our columns manually, rather than it automatically discovering it from the head.

why is having an iterable [] instead of None forcing you to define columns manually?

[alikleit]

My bad!

skip_row=0 solved my problem! - had it on 1 by default. thought that skip_row is not related to the header and header would be discovered when setting it to 'None'.

Thanks for the quick response though appreciated!! 😄