Apache Spark & Scala certification Training Gurgaon,Delhi - Gyansetu

Overview of Hadoop & Apache PySpark 

This will cover the introduction to Hadoop ecosystem with a little insight to Apache Hive.

Introduction to Big Data and Data Science–

• Learn about big data and see examples of how data science can leverage big data
• Performing Data Science and Preparing Data – explore data science definitions and topics, and the process of preparing data

1) Introduction

2) PySpark Installation

3) PySpark & its Features

    > Speed

    > Reusability

    > Advance Analytics

    > In Memory Computation

    > RealTime Stream processing

    > Lazy Evaluation

    > Dynamic in Nature

    > Immutability

    > Partitioning

4) PySpark with Hadoop

5) PySpark Components

6) PySpark Architecture

1) RDD Overview

2) RDD Types (Ways to Create RDD):

• Parallelized RDDs(PairRDDs) > RDD from Collection Objects : Example: map() & flatMap() function so that we able to perform groupByKey() & others....
• RDD from External Datasets (CSV,JSON,XML.....)
  • rowRDD
  • schemaRDD (Adding schema in RDD particularly Dataframes)

3) RDD Operations:

• Basic Operations like map() & flatMap() to convert into pairRDD.
• Actions
• Transformations : 

Passing Functions to PySpark

Working with Key-Value Pairs (pairRDDs)

SuffleRDD : ShuffleOperations Background & Performance Impact

RDD Persistence & Unpersist

PySpark Deployment :

    1) Local

    2) Cluster Modes:

        a) Client mode

        b) Cluster mode

Shared Variables:

    1) Broadcast Variables

    2) Accumulators

Launching PySpark Jobs from Java/Scala

Unit Test Cases

PySpark API :

    1) PySpark Core Module

    2) PySpark SQL Module

    3) PySpark Streaming Module

    4) PySpark ML Module

Integrating with various Datasources (SQL+ NoSQL + Cloud)

Dataframes: 

1) Overview

• An introduction to PySpark framework in cluster computing space.
• Reading data from different PySpark sources.
• PySpark Dataframes Actions & Transformations.
• PySpark SQL & Dataframes
• Basic Hadoop functionality
• Windowing functions in Hive
• PySpark Architecture and Components
• PySpark SQL and Dataframes
• PySpark Data frames
• PySpark SQL
• Running PySpark on cluster
• PySpark Performance Tips

2) Creating Dataframes using PySparkSession

3) Untyped Tranformations

4) Running SQL Queries Programatically

5) Global Temporary View

6) Interoperating with RDDs:

1. Inferring Schema using Reflection
2. Programatically specifying the Schema

7) Aggregations:

1. Untyped User-Defined Aggregate Functions
2. Type-Safe User-Defined Aggregare Functions

1) Generic LOAD/Save Functions

• Manually Specifying Options
• Run SQL on files directly (avro/parquet files)
• Saving to Persistent Tables ( persistent tables into HiveMetastore)
• Bucketing, Sorting & Partitioning (  repartition() &  coalesc() )

 2) Parquet Files

1.  Loading Data Programatically
2.  Partition Discovery
3.  Schema Merging
4.  Hive Metastore Parquet Table Conversion 

• Hive Parquet/Schema Reconcialation
• Metadata Refreshing

5. Configuration

3) ORC Files

4) JSON Datasets

5) Hive Tables:

• Specifying storage format for Hive Tables
• Interacting with various versions of Hive Metastore.

a) JDBC to other databases (SQL+ NoSQL databases)

b) Troubleshooting

c) PySpark SQL Optimizer

d) Transforming Complex Datatypes

e) Handling BadRecords & Files 

f) Task Preemptionh Concurrency  for Hig

Optional>>Only for DataBricks Cloud

g) Handling Large Queries in Interactive Flows

• Only for DataBricks Cloud (Using WatchDog)

h) Skew Join Optimization

i) PySpark UDF Scala 

j) PySpark UDF Python

k) PySpark UDAF Scala

l) Peformance Tuning/Optimization PySpark Jobs:

• Caching Data in Memory
• Other Configurations like:

                      >maxPartitionBytes

                      >openCostInBytes

                     >broadcastTimeout

                     >autoBroadcastJoinThreshold

                     >partitions

• BHJ (BroadcastHashJoin)
• Serializers

m) Distributed SQL Engine: (Accessing PySpark SQL using JDBC/ODBC using Thrift API)

1.  Running the Thrift JDBC/ODBC server
2.  Running the PySpark SQL CLI (Accessing PySpark SQL using SQL shell)

n) PySpark for Pandas using Apache Arrow:

1. Apache Arrow in PySpark (To convert PySpark dataframes into Python dataframes)
2. Conversion to/from Pandas
3. Pandas UDFs

o) PySpark SQL compatibility with Apache Hive

p) Working with DataFrames Python & Scala

q) Connectivity with various DataSources

a) Introduction to DataSets

b) DataSet API - DataSets Operators

c) DataSet API - Typed Tranformations

d) Datasets API - Basic Actions

e) DataSet API - Structured Query with Encoder

f) Windowing & Aggregation DataSet API

g) DataSet Checking & Persistence

h) DataSet Checkpointing

i) Datsets Vs Dataframes Vs RDDs

a) Overview

b) Example to demonstrate DStreams

c) PySpark Streaming Basic Concepts:

1. Linking
2. Initialize Streaming Context
3. Discretized Streams (DStreams)
4. Input DStreams & Receivers
5. Transformations on DStreams
6. Output Operations on DStreams : print(), saveAsTextFile(), saveAsObjectFiles(), saveAsHadoopFiles(), foreachRDD()
7. Dataframe & SQL Operations on DStreams (Converting DStreams into Dataframe)
8. Streaming Machine Learning on streaming data
9. Caching/Persistence
10. Checkpointing
11. Accumulators, Broadcast Variables, and Checkpoints
12. Deploying PySpark Streaming Applications.
13. Monitoring PySpark Streaming Applications

d) Performance Tuning PySpark Streaming Applications:

1. Reducing the Batch Processing Times
2. Setting the Right Batch Interval
3. Memory Tuning

e) Fault Tolerance Semantics

f) Integration:

1. Kafka Integration
2. Kinesis Integration
3. Flume Integration

g) Custom Receivers : creating client/server application PySpark Streaming.

a) Overview

b) Example to demonstrate GraphX

c) PropertyGraph : Example Property Graph

d) Graph Operators:

1. Summary List of Operators
2. Property Operators
3. Structural Operators
4. Join Operators
5. Neighbourhood Aggregation:

             >Aggregation Messages (aggregateMessages)

            >Map Reduce Triplets Transtion Guide (Legacy)

            >Computing Degree Information

             >Collecting Neighbours

6. Caching & Uncaching

e) Pregel API

f) Graph Builders

g) Vertex & Edge RDDs:

1. VertexRDDs
2. EdgeRDDs

h) Optimiation Representation

i) Graph Algorithms

1. PageRank
2. Connected Components
3. Triangle Counting

j) Examples

k) GraphFrames & GraphX

• How to manage & Monitor Apache Spark on Kubernetes
• Spark Submit Vs Kubernetes Operator
• How Spark Submit works with Kubernetes
• How Kubernetes Operator for Spark Works.
• Setting up of Hadoop Cluster on Docker
• Deploying MR, Sqoop & Hive Jobs inside Hadoop Dockerized environment.

• Building & running applications using PySpark API.
• PySpark SQL with mySQL (JDBC) source :
• Now that we have PySpark SQL experience with CSV and JSON, connecting and using a mySQL database will be easy. So, let’s cover how to use PySpark SQL with Python and a mySQL database input data source.
• Overview
• We’re going to load some NYC Uber data into a database. Then, we’re going to fire up PySpark with a command line argument to specifiy the JDBC driver needed to connect to the JDBC data source. We’ll make sure we can authenticate and then start running some queries.

** will also be going through some of the concepts like caching and UDFs.

A real-world business case built on top of PySpark API. 

1) Web Server Log Analysis with Apache PySpark – use PySpark to explore a NASA Apache web server log.

2) Introduction to Machine Learning with Apache PySpark – use PySpark’s MLB Machine Learning library to perform collaborative filtering on a movie dataset

3) PySpark SQL with New York City Uber Trips CSV Source: PySpark SQL uses a type of Resilient Distributed Dataset called DataFrames which are composed of Row objects accompanied by a schema. The schema describes the data types of each column. A DataFrame may be considered similar to a table in a traditional relational database.

Methodology

We’re going to use the Uber dataset and the PySpark-CSV package available from PySpark Packages to make our lives easier. The PySpark-CSV package is described as a “library for parsing and querying CSV data with Apache PySpark, for PySpark SQL and DataFrames” This library is compatible with PySpark 1.3 and above.

 PySpark SQL with New York City Uber Trips CSV Source

Software Required :

·         VMWare Workstation

·         Ubuntu ISO Image setup on Virtual Environment(VMWare)

·         Cloudera Quickstart VM (version : 5.3.0)

·         Putty Client

·         WinSCP

·         Hadoop software version 2.6.6

·         PySpark 2.x

Apache Spark is an open source big data processing framework built around speed, ease of use, and sophisticated analytics. It enables applications in Hadoop clusters to run up to 100 times faster in memory and 10 times faster even when running on disk. This is making it an inevitable technology and everyone who wants to stay in big data engineering is keep to become an expert in Apache Spark.

The demand for Apache spark is gaining momentum in current job market. There is huge shortage of Analysts with hands on experience on Apache Spark. This is also the main reason they are drawing handsome salaries.

Apache Spark will help you understand how Spark executes in-memory data processing & how Spark Job runs faster then Hadoop MapReduce Job. Course will also help you understand the Spark Ecosystem & it related APIs like Spark SQL, Spark Streaming, Spark MLib, Spark GraphX & Spark Core concepts as well. This course will help you to understand Data Analytics & Machine Learning algorithms applying to various datasets to process & to analyze large amount of data.

Spark Ecosystem:

• Understanding Scala in its implementation & in Apache Spark
• OOPs concepts in Scala Programming language.
• How to build Scala  & Spark using SBT.
• Scala & Spark installation .
• Spark operations on Spark Shell.
• How to submit Scala in Spark environment.
• Spark Driver & its related Worker Nodes.
• Spark + Flume Integration.
• Setting up Data Pipeline using Apache Flume, Apache Kafka & Spark Streaming
• Spark RDDs.
• Spark RDDs Actions & Transformations.
• Spark SQL : Connectivity with various Relational sources & its convert it into Data Frame using Spark SQL.
• Spark Streaming
• Understanding role of RDD.
• Spark MLib : Creating Classifiers & Recommendations systems using MLib .
• Spark Core concepts : Creating of RDDs: Parrallel RDDs, MappedRDD, HadoopRDD, JdbcRDD.
• Spark Architecture & Components.
• Spark Structure Streaming.
• Spark SQL experience with CSV , XML & JSON.
• Reading data from different Spark sources.
• Spark SQL & Dataframes.
• Implementing Accumulators & Broadcast variables for Performance tuning.
• AWS Cloud
• Deploying BIG Data application in Production environment using Docker & Kubernetes

Our Big Data Experts have realized that Learning Hadoop standalone doesn’t qualify candidates to clear the interview process. Interviewers demand and expectations from the candidates are more nowadays. They expect proficiency in advanced concepts like-

• Expertise in PySpark/ Scala-Spark
• Real Time Data Storage inside Data Lake (Real Time ETL)
• Big Data Related Services on AWS Cloud
• Deploying BIG Data application in Production environment using Docker & Kubernetes
• Experience in Real Time Big Data Projects

All the advanced level topics will be covered at Gyansetu in a classroom/online Instructor led mode with recordings.

Knowledge of Python/ Scala, SQL is good to start Spark Training in Gurgaon. However, Gyansetu offers a complementary instructor led course on Python/ Scala, SQL before you start Spark course.

Gyansetu is providing complimentary placement service to all students. Gyansetu Placement Team consistently work on industry collaboration and associations which help our students to find their dream job right after the completion of training.

• Our placement team will add Big Data Hadoop, PySpark skills & projects in your CV and update your profile on Job search engines like Naukri, Indeed, Monster, etc. This will increase your profile visibility in top recruiter search and ultimately increase interview calls by 5x.
• Our faculty offers extended support to students by clearing doubts faced during the interview and preparing them for the upcoming interviews.
• Gyansetu’s Students are currently working in Companies like Sapient, Capgemini, TCS, Sopra, HCL, Birlasoft, Wipro, Accenture, Zomato, Ola Cabs, Oyo Rooms, etc.

• Gyansetu trainer’s are well known in Industry; who are highly qualified and currently working in top MNCs.
• We provide interaction with faculty before the course starts.
• Our experts help students in learning Technology from basics, even if you are not good in basic programming skills, don’t worry! We will help you.
• Faculties will help you in preparing project reports & presentations.
• Students will be provided Mentoring sessions by Experts.