DatasetGenerationError. Failed to parse string: as a scalar of type double

why it decided to encode it as float instead

I think that probably made that’s decision based on the first few lines…:sweat_smile:
It seems that external libraries (Pandas and PyArrow) are used for parsing CSV and JSON, and that’s probably how it works. It seems that things like on_bad_lines=“skip” are also completely thrown over to them.

1 Like