this approach relies on the fx:command
property to transform the parquet file into csv
justin@X~/github/sparql.anything$ bash -c "python3 -c \"import pandas as pd ; pd.read_parquet(\\\"some.parquet\\\").to_csv(\\\"/dev/stdout\\\")\""
,name,age
0,bob,45
1,fred,3
then running the query below like this: