GettingStarted

This document is for those who …

• …tried R but are not used to Hadoop and Map/Reduce.
• …tried using SQL but not proficient with it and neither with working by typing in Command Lines.
• …want to keep or at least similarly maintain the current R-using analysis style but not fussing much about distributed environments and massive volumes.
• …want to be quickly immersed and experienced in distributed or big data analysis.

Contents

• Installation required for operating RHive, configuring settings and simple troubleshooting.
• RHive examples
• Those R users who want to use RHive for big data analysis can use the user guide to learn basic instructions and accomplish fundamentals of analysis.

---

RHive – GettingStarted
This document is written in simple, intuitive instructions for users new to RHive.

What is RHive?
R language + Hive (Extension software of Hadoop that allows using SQL syntax to approach files stored in Hadoop and Map/Reduce task)
RHive is an R package that makes distributed data processing and big data analysis easy by using R language’s syntax to approach R or use it for Map/Reduce framework.
As a composite of R package and Hive, users can simultaneously use R language and Hive and also supports use of R language and SQL syntax for big data analysis.

Things to know before using RHive

• Basic User
  • R language (GNU-R) syntax
    • R is an open source language for statistics and analysis(can be used for open source’s SAS, SPSS, MATLAB etc.)
  • Basic Concepts of Map/Reduce
    • Basic concepts of Map/Reduce, which is the most widely used method of processing distributed data.
  • Basic Concepts of Hadoop
  • Basic Concepts of Hive
• Advanced User
  • Hive SQL syntax, concepts and the workings of UDF, UDAF
  • Hadoop file system(HDFS)
  • Advanced R syntax

Introduction to RHive
Understanding RHive

• Enables using big data and map/reduce work to be dealt without making implementations for them but using only R and SQL syntax.
• With but simple tasks in R, aggregating, preprocessing, and basic statistical analyzing become easy.
• There’s no need to strain to obtain a complete understanding of Map/reduce.
• Reference : Hive UDF

Pros

• With nothing else required but knowledge of R and SQL syntax, anyone can process big data.

Cons

• Requires at least an inkling of understanding of map/reduce and Hive.
  • Analysts virtually know nil about this.
• Must know about SQL syntax.
  • Quite a significant population of analysts do not know about this.
• Debugging is difficult (the root of this problem lies not in RHive, but in distributed environment).