JSON streaming for Spark or Dataproc

This project demonstrates how to stream JSON data for: Google Cloud Dataproc or a local Spark instance. The script reads the JSON data, infers the schema, and then creates/loads a Hive table with the data for quick insight and integration.

Requirements

• Python 3.6 or higher
• pyspark
• google-cloud-storage (only required for the data_to_dataproc.py script)

To install the required packages, you can run:

pip install pyspark google-cloud-storage

Configuration

Update the config.json file with the appropriate settings for your use case. This file contains separate configurations for Google Cloud Storage and local Spark instances, as well as the database and table names.

Usage

Run the appropriate script based on your desired instance:

For Google Cloud Dataproc

submit a new job for your cluster: gcloud dataproc jobs submit pyspark app/data_to_dataproc.py --cluster=<CLUSTER-NAME> --region=<REGION> --py-files= app/common.py --files=config.json

For a local Spark instance

simply run: python app/data_to_spark.py: