big data: big sql application development with biginsights
DESCRIPTION
IBM, Hadoop, Big Data, BigInsights, application development, SQL, Big SQL, Web console, lifecycleTRANSCRIPT
Developing, publishing and deploying your first Big SQL application with InfoSphere BigInsights
Cynthia Saracco
Senior Solution Architect
IBM Silicon Valley Lab
2 © 2013 IBM Corporation
Big Data Application Ecosystem
Eclipse
App library
MapReduce, "
Text Analytics
Query
App Development
• Code application program, and generate
associated App
• Deploy Apps to Enterprise ManagerApp
Development
Publish
Data integration scenario:
Pre-defined work flows simplify
loading data from various
sources
•Work flows can be configured,
deployed, executed and
scheduled
Development tooling:
•Text analytics
•MapReduce
•Query languages
• . . .
Application scenarios (web log,
email, social media, �):
• Samples provide starting
point, speed time to value
Big Data Web Console
3 © 2013 IBM Corporation
BigInsights Applications Catalog (Web Console)
� Browse available applications � Manage and deploy applications (administrators only) � Execute (or schedule execution of ) a deployed application � Monitor job (application) status� Link or chain applications for sequential execution
4 © 2013 IBM Corporation
Overview of Application Development Lifecycle
� Configure your Eclipse environment (one-time set up) � Develop your application using BigInsights tools � Test your application � Package and publish your application � Deploy your application on the cluster
5 © 2013 IBM Corporation
Configure your Eclipse environment
� One-time set up � Download and install BigInsights tools (Eclipse plug-ins)
– Welcome tab of BigInsights Web console includes pre-req info, download &
installation instructions
6 © 2013 IBM Corporation
Develop your application – Big SQL example
� Open the BigInsights perspective in Eclipse� Create a BigInsights project
7 © 2013 IBM Corporation
Develop your application – Big SQL example (cont’d)
� Create a BigInsights program with a SQL script (file)
8 © 2013 IBM Corporation
Develop your application – Big SQL example (cont’d)
� Populate your SQL file with the desired code
create table if not exists media_csv
(id integer not null,
name varchar(50),
url varchar(50),
contactdate string)
row format delimited
fields terminated by ','
stored as textfile;
load hive data local inpath
'/home/biadmin/sampleData/DBMS/RDBMS_data.csv'
-- overwrite
into table media_csv;
9 © 2013 IBM Corporation
Test your application
� Run your application from Eclipse
10 © 2013 IBM Corporation
Publish your application to the BigInsights catalog
� Package and publish your application from Eclipse � Specify application name, workflow requirements, etc.
11 © 2013 IBM Corporation
Deploy your application on the cluster
� Access the Applications tab of the Web console� “Manage” the published applications � Locate your new application and deploy it � Create credentials store file in DFS (if needed) -- see next chart � Optionally, execute the application after it’s been deployed
12 © 2013 IBM Corporation
Run your Big SQL application on the cluster
13 © 2013 IBM Corporation
Upgrade your application (optional)
� Satisfy evolving business requirements, improve flexibility – Example: add input parm(s)
� Modify SQL code and re-package, re-publish, re-deploy
create table if not exists $TABLE(id integer not null,
name varchar(50),
url varchar(50),
contactdate string)
row format delimited
fields terminated by ','
stored as textfile;
load hive data local inpath '$FILE' -- overwrite
into table $TABLE;
14 © 2013 IBM Corporation
Upgrade your application (cont’d)
� Re-publish your application from Eclipse � Adjust workflow specs for input parm(s)
15 © 2013 IBM Corporation
Summary
� Eclipse tools simplify big data application development for BigInsights– Wizards
– Context-sensitive help
– Oozie workflow generation
– Built-in test environment
– Etc.
� Application catalog provides easy way to locate and launch apps of interest– Developers use Eclipse tools to package/publish their applications to this
catalog
– Application upgrades easily managed