Batch Load Parquet Files to Lakehouse Using Copy Command

This guide will help you use scripts and the Singdata Lakehouse SQLLine command-line tool to import data from a large number of public URL Parquet files (e.g., New York City Taxi and For-Hire Vehicle Data) into Singdata Lakehouse.

1. download_raw_data.sh, download Public URL Parquet files to the local machine. This article uses NYC Taxi Data as an example, which includes over 450 files and 60GB of data (Parquet format).

2. initialize_database.sh, create the schema and tables in Singdata Lakehouse using the Lakehouse SQLLine command.

3. Import taxi and FHV data, use R to convert Parquet files to CSV files, and use the Lakehouse SQLLine copy command to import data from local CSV files into Singdata Lakehouse tables, supporting batch import of multiple files.

Operation Guide

Script download address: https://github.com/yunqiqiliang/nyc-taxi-data-clickzetta

1. Install ClickZetta SQLLine

2. Install R

From CRAN

Note that R used to be optional for this repo, but is required starting with the 2022 file format change. The scripts use R to convert Parquet files to CSV before loading into Postgres. There are other ways to convert from Parquet to CSV that wouldn't require R, but I found that R's arrow package was faster than some of the other CLI tools I tried

3. Download raw data

./download_raw_data.sh

while done modify download_raw_data.sh and run again to download data from 202212: wget -i setup_files/raw_data_urls_new.txt -P data/ -w 2

4. Initialize database and set up schema

./initialize_database.sh

5. Import taxi and FHV data

./import_yellow_taxi_trip_data.sh

./import_green_taxi_trip_data.sh

./import_fhv_taxi_trip_data.sh

./import_fhvhv_trip_data.sh

Note that the full import process might take several hours or possibly even over a day depending on computing power

Schema

• trips table contains all yellow and green taxi trips. Each trip has a cab_type_id, which references the cab_types table and refers to one of yellow or green
• fhv_trips table contains all for-hire vehicle trip records, including ride-hailing apps Uber, Lyft, Via, and Juno
• fhv_bases maps fhv_trips to base names and "doing business as" labels, which include ride-hailing app names
• nyct2010 table contains NYC census tracts plus the Newark Airport. It also maps census tracts to NYC's official neighborhood tabulation areas
• taxi_zones table contains the TLC's official taxi zone boundaries. Starting in July 2016, the TLC no longer provides pickup and dropoff coordinates. Instead, each trip comes with taxi zone pickup and dropoff location IDs
• central_park_weather_observations has summary weather data by date

Load Script Example

Full Script

#!/bin/bash

fhv_schema="(dispatching_base_num, pickup_datetime, dropoff_datetime, pickup_location_id, dropoff_location_id, legacy_shared_ride_flag, affiliated_base_num)"

for parquet_filename in data/fhv_tripdata*.parquet; do
  echo "`date`: converting ${parquet_filename} to csv"
  ./setup_files/convert_parquet_to_csv.R ${parquet_filename}

  csv_filename=${parquet_filename/.parquet/.csv}
  
  sh ../sqlline_cz/sqlline properties ../sqlline_cz/clickzetta.properties -e "set copy.csv.with.header=false;set copy.csv.skip.header=true; copy fhv_trips_staging from '${csv_filename}';"

  echo "`date`: finished raw load for ${csv_filename}"

  rm -f $csv_filename
  echo "`date`: deleted ${csv_filename}"
done;