Join GitHub today
GitHub is home to over 36 million developers working together to host and review code, manage projects, and build software together.Sign up
The feather example from docs segfault with pyarrow 0.8.0 #24767
The example from the docs succeeds to write and read the feather file, but the read-in dataframe is corrupted.
On general displaying of the
The above is with latest master of pandas, and pyarrow 0.8.0 installed from conda-forge (quite old, but still in the supported range). Still need to try with a newer version of pyarrow.
The above was in my development environment (where I apparently still has an old pyarrow), but I now created a clean new env just with installing pyarrow and pandas, and can confirm the issue. With pyarrow 0.11 it works fine, but reading the file with pyarrow 0.8.0 gives a segfault.
Of course, it can also be a bug in the old version of pyarrow. But the question is then if we need to guard users against it (if there is some way to detect the invalid data coming from pyarrow).
We bumped fastparquet to its latest.…
________________________________ From: gfyoung <firstname.lastname@example.org> Sent: Sunday, January 20, 2019 18:05 To: pandas-dev/pandas Cc: Subscribed Subject: Re: [pandas-dev/pandas] The feather example from docs segfault with pyarrow 0.8.0 (#24767) I think that seems reasonable to squeeze in at this time. Should we also bump up the min version of fastparquet (similar to #23482<#23482>), or is that not necessary? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#24767 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ABQHItbhPjJBj1eyWsi0fOclSBQ59etHks5vFQQ5gaJpZM4Z-5BB>.