You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Additional information
This makes it impossible to perform an export to HDF5 or other operations that do not support the object type. This is especially blocking when loading CSV files.
Should be relevant to #1568.
The text was updated successfully, but these errors were encountered:
Vaex does not officially support object columns or columns of mixed types. Please cast the data to a single type before passing it to vaex. This is kind of like a database behavior sort of.
Reading CSVs should work. There the behavior is a bit different. If you encounter an issue, please made a small CSV file that reproduces the issue.
I do think this would be a nice feature. But since dtype=object data is in memory anyways, I think it's easy to do this beforehand. It's unlikely to be big data, nor will it work multithreaded.
Thanks for the report, lets keep this open an a low priority feature for now!
@maartenbreddels
same issue here. i tried to left join 2 tables (so have None value in some cell) which the join key is calculated string column, but then it automatically converted to object type and can't export to hdf5 file.
Description
Trying to convert an "object" column to string fails.
Example code:
Exception:
Software information
import vaex; vaex.__version__)
:Additional information
This makes it impossible to perform an export to HDF5 or other operations that do not support the object type. This is especially blocking when loading CSV files.
Should be relevant to #1568.
The text was updated successfully, but these errors were encountered: