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
It can be quite slow working with the existing metadata since it is in csv format - my first step is often to save a copy in a binary format, but this isn't an ideal workflow.
I'm curious whether sharing a binary format (feather/parquet) would be possible, either in addition to or replacing the CSVs.
This would also eliminate the need for specifying dtypes on load, which can currently cause confusing errors if not attended to.
The text was updated successfully, but these errors were encountered:
It can be quite slow working with the existing metadata since it is in csv format - my first step is often to save a copy in a binary format, but this isn't an ideal workflow.
I'm curious whether sharing a binary format (feather/parquet) would be possible, either in addition to or replacing the CSVs.
This would also eliminate the need for specifying dtypes on load, which can currently cause confusing errors if not attended to.
The text was updated successfully, but these errors were encountered: