c929b60
1
2
3
4
5
6
import pandas as pd df = pd.read_parquet('mscoco.parquet') first_caps = df.drop_duplicates(subset='URL', keep='first') first_caps.to_parquet('mscoco-1st-caption.parquet')
