Read the first couple of rows in Parquet file stored in S3

parquet-head

#!/usr/bin/env bash

profile=default
limit=100

while [ -n "$1" ]; do
    case $1 in
        --profile )
            shift
            profile=$1
            ;;

        --limit )
            shift
            limit=$1
            ;;

        * )
            inputfile=$1
            ;;
    esac
    shift
done

inputfile=${inputfile/s3\:\/\//}
key=${inputfile#*/}
bucket=${inputfile%%/*}

if [[ -z "$key" ]] || [[ -z "$bucket" ]]
then
    echo "
        Invalid S3 file path
    "
    exit 1
fi

aws --profile ${profile} s3api select-object-content \
    --bucket ${bucket} \
    --key ${key} \
    --expression "SELECT * FROM S3Object LIMIT ${limit}" \
    --expression-type SQL \
    --input-serialization 'Parquet={}' \
    --output-serialization 'JSON={}' \
    /dev/stdout