Error loading dataset
#4
by
alexmsrh
- opened
Hi,
I get an error when loading the dataset usingdatasets.load_dataset("zhengyun21/PMC-Patients")
Generating train split: 0 examples [00:00, ? examples/s]Failed to convert pandas DataFrame to Arrow Table from file 'HF_CACHE/hub/datasets--zhengyun21--PMC-Patients/snapshots/28d8836518f86d4f1e6358ea8ec09977023e5766/PMC-Patients-V2.json' with error <class 'pyarrow.lib.ArrowInvalid'>: ("Could not convert 'year' with type str: tried to convert to double", 'Conversion failed for column age with type object')
I'm using datasets 3.3.2. Downgrading to 3.2.0 does not solve it.
Would you have any idea how to solve this?
Hi, we think the error may be caused by that we upload a new version of PMC-Patients in JSON format. We suggest you simply download the file and load it with python. Note that the new version is not necessarily the superset of the old version since the PMC release may remove some articles.