Memo Parquet

En utilisant pandas

import pandas as pd
import s3fs

s3 = s3fs.S3FileSystem()
df = pd.read_parquet('my-bucket/my-file.parquet', filesystem=s3, engine='pyarrow')

En utilisant pyarrow

import s3fs
import pyarrow.parquet as pq

s3 = s3fs.S3FileSystem()
df = pq.read_table('my-bucket/my-file.parquet', filesystem=s3)