import pandas as pd
from pathlib import Path

# Specify the folder path containing parquet files
folder_path = "/Users/fredericlegrand/Documents/Code/Python tests/temp/data/chunk-000" 

# Iterate through all parquet files in the folder
for file in Path(folder_path).glob("*.parquet"):
    # Read parquet file
    df = pd.read_parquet(file)
    
    # Slice the first 6 items from the arrays in "action" and "observation.state"
    df["action"] = df["action"].apply(lambda x: x[:6])
    df["observation.state"] = df["observation.state"].apply(lambda x: x[:6])
    
    # Overwrite the original file with the modified data
    df.to_parquet(file)
    
    print(f"Processed {file.name}")

print("All files processed.")