fix_parquet.py

import pandas as pd
import glob

parquet_files = glob.glob('data/chunk-000/episode*.parquet')


for file in parquet_files:

    df = pd.read_parquet(file)
    file_index = int(file.split("_")[-1].split(".")[0])
    print(df["observation.state"])