Hive Setup

Hive Setup is quite straightforward could be achieved by following just 1,2,3 steps. Below is brief description how to setup Hive.

Prerequisites
1. Java 6
2. Installed Hadoop2.x as explain in section Hadoop Setup
3. Cygwin in case of Windows OS

Download
1. Download tar from link http://hive.apache.org/releases.html
2. $ tar -xzvf hive-x.y.z.tar.gz
3. Extract into /Users/nitin/software/hive-a.b.c

Setup Environment
1. $ export HIVE_HOME=/Users/application/apache-hive-0.13.0
2. $ export PATH=$HIVE_HOME/bin:$PATH
3. Add it into your bash profile so you can avoid setting again and again
4. Launch the Hive shell as below
% hive
hive>
Hive could be launch using hive command or we could also use the -e option run from command inline.
By default Hive keeps the default database

hive&amp;amp;gt; show databases;
Or $hive -e 'show databases'
OK
Default

HiveQL is not case sensitive so ‘show databases’ works also.
Below command use to display the list of tables in current databases which is ‘default’

Hive&amp;amp;gt; show tables;
OK

Tables
A Hive table’s data stored in HDFS filesystem and associate with schema stored as metadata in Metastore locally.
Table could be basically two types
1. Managed table: Hive managed the data stored in HDFS and if table got drop then data would also removed from HDFS.
2. External table: External table will map already existing data in HDFS filesystem and if table dropped data will still available in HDFS filesystem
Create Table:
Create file with value
Employee.txt
================
name1,address1
name2,address2

Hadoop fs –copyFromLocal employee.txt /tmp/employee/
CREATE EXTERNAL TABLE IF NOT EXISTS employee (NAME STRING, ADDRESS STRING)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ','
LINES TERMINATED BY 'n'
LOCATION '/tmp/employee';
hive&amp;amp;gt;select * from employee;
OK
name1 address1
name2 address2

Reference
Hadoop Essence: The Beginner’s Guide to Hadoop & Hive