Both machine_new and machine_old were running pandas 1.x. Just upgraded machine_new to pandas 2.x, the pandas data frame is saved by pickle.dump() on machine_new, the file cannot be read by pickle.load() on machine_old.
pickle.format_version on machine_old and machine_new are both 4.0.
The specific error msg is: AttributeError: Can't get attribute '_unpickle_block' on <module 'pandas._libs.internals' from '/anaconda3/lib/python3.8/site-packages/pandas/_libs/internals.cpython-38-x86_64-linux-gnu.so'>
Using pd.read_pickle(f) instead of pickle.load() does not help, as suggested in AttributeError: Can't get attribute '_unpickle_block'.
Other than upgrading/downgrading pandas, is there a way to solve this?
Thanks.
