README.md

Introduction

The repository contains code snippets and explanation of working with different hadoop file formats i.e. ORC, Parquet and Avro, Spark local and standalone architecture

File Structure

Local_Mode_Architecture: Describes the local mode architecture, different ways to use local mode and memory allocation

  Tip: To try different local mode settings, changes should be made in conf file(located in spark folder), before running file.

File_formats: Describes the avaliable options in spark to read different file formats and evaluation of each option and file format

  Tip: There are few external packages for data formats, avaliable as jar files in Maven repository.Should be downloaded and saved in jars folder(located in spark folder), before running file.

Standalone_Mode_Architecture: Describes setting up spark cluster in standalone mode in single machine, setting up environment and memory allocation

  Tip: Configuration should be set up in spark-env.cmd(Windows) and Masters-slaves are initiated by powershell commands(powershell_command.ps1) before running file.

Tools and Environment

spark-3.0.1-bin-hadoop2.7, pyspark library, Jupyter Notebook, Windows 10

How to use the project

To learn some tool, it is always better to do hands on. Through this project, I have explored above mentioned three areas of Spark, learnt through the way by challenging myself to understand and question all related 'whys'. I have gathered information from multiple sources mainly from stackoverflow and Apache documentation to clarify doubts. There may be gaps in information which I may not be able to capture due to my prior familiarity with the topics itself. But this serves as a good supplement/ revision guide for Spark learners and users.