Saturday, 26 October 2013

Oracle Apache Hadoop Hive ODBC Driver with the Oracle BI 11g Administration tool


OBIEE 11.1.1.7 introduces capabilities to integrate with Hadoop sources using the Hive ODBC interface. I believe it is a significant step forward which enhances the data integration abilities of the OBIEE.
The purpose of this blog is to share a document from oracle support which describes how to use the Oracle Apache Hadoop Hive ODBC Driver with the Oracle BI Administration tool for the purpose of creating Oracle BI metadata from Hadoop Data sources. I have used Sample Apps screen snapshots within this blog.

Pre-requisites – ODBC Driver

Installers for 32-bit and/or 64-bit Windows ODBC drivers to be used with the Oracle BI Administration tool are a pre-requisite for ODBC connection.

The following zip files are available to download from Oracle Support.

Win32Hive.zip - For 32 bit Operating System
Win64Hive.zip – For 64 bit Operating System 

Installation Steps for Windows

  1. Backup the Windows registry file before beginning installation. Extract the zip files consistent with the platform on which the BI Administration tool is installed.

  1. Run install.bat from the extracted sub-directly. Please note that the installation needs to be run as a Windows local administrator.

ODBC Data Source

After installation of the driver on Windows, you may now configure ODBC Data Sources to the particular Hive/Hadoop system you plan to connect to.






On UNIX Server once installation of ODBC driver is done; please make sure that odbc.ini should have been updated with the Hive ODBC information ORACLE_INSTANCE/bifoundation/OracleBIApplication/coreapplication/setup/odbc.ini


The detail about ODBC ‘Hive’ connection is as below

 
Database and Connection Pool for Hadoop

From the BI Administration tool, import your data source using the ‘ODBC 2.0’ connection type. After creating metadata and before deploying the RPD to the BI Server, you will need to change physical database type to ‘Apache Hadoop’ and change the connection pool to a valid DSN name on the BI Server as shown below.
  • Database Definition for Apache Hadoop 


 
  •  Connection Pool for Hadoop

 


No comments:

Post a Comment