We want to create a loader for Iceberg. The objective is to extract a CSV file using

tap-csv

and put the data as parquet file into an

S3 bucket

which is connected to our Iceberg datalake. Should we take

target-s3

as a reference and customize it to follow this: https://iceberg.apache.org/spec/#parquet or should we use

pyspark

to write files to our data lake? Not sure how to approach this. I'm really looking for a direction from the community.