- Install hive
1. download hive-0.12.0.bin.tar.gz
2.#tar -xzvf hive-0.12.0.bin.tar.gz
3.add the bin dir to PATH in ~/.bashrc
4.#source ~/.bashrc
The dir structure of hive-0.12.0.bin likes the following:
lib/ : contains JARs., which implement a particular subset of Hive's functionality.
bin/: contains executable scripts that launch various Hive service, including the hive command-line interface(CLI)
hcatalog/: metastore service related procedures.
conf/: configuration files
examples/: some example script
Thrift service provides remote access from other processes.
Hive uses a built-in Derby SQL server, which provides limited, single-process metadata storage.
start hive
#hive //start hive CLI
#hive --help
The list of hive service is :
you can start a service using command like
#hive --service service-name
#hive //default run CLI service
look for help for CLI
#hive --help --service cli
- Variables and Properties
In CLI, set and display variables and properties using set
#hive --define foo=bar
#hive
hive> set foo;
hive>set;
hive>set -v;
hive>set env:HOME;
hive>set foo=xbar;
- Basic commands
#hive --hiveconf hive.cli.print.current.db=true
#hive -e "select * from mytable limit 3";
#hive -S -e "select * from mytable limit 3" > /tmp/myquery;
execute hql script
#hive -f /path/to/withqueries.hql
or
hive> source /path/to/withqueries.hql
shell execution
hive> ! shell-commands;
hive> ! pwd;
note: Don't invoke interactive commands that require user input. shell 'pipes' don't work and neither do file 'globs'.
Hadoop dfs commands from inside hive
hive> dfs -ls /;
hive> dfs -help;
Comments in Hive Scripts
--the is a comment line
Query Column Headers
hive>set hive.cli.print.header=true;
Start Hive Server
#hive --service hiveserver & //run hiveserver in daemon mode
or start hiveserver2
#hiveserver2
#netstat -nl |grep 10000 //check whether the server is running
References:
https://cwiki.apache.org/confluence/display/Hive/Home