You are browsing a read-only backup copy of Wikitech. The live site can be found at wikitech.wikimedia.org
Beeline is the command line shell that ships with HiveServer2, which was introduced in Hive 0.11. The original Hive CLI is being deprecated in favor of Beeline, and the Analytics team at WMF now supports Beeline as a Hive client. The original Hive client will continue to exist on the stat boxes until it's deprecated fully, but we encourage moving to beeline.
SSH into stat1002/1004. Run
You should get a prompt like:
madhuvishy@stat1004:~$ beeline Picked up JAVA_TOOL_OPTIONS: -Dfile.encoding=UTF-8 Picked up JAVA_TOOL_OPTIONS: -Dfile.encoding=UTF-8 Picked up JAVA_TOOL_OPTIONS: -Dfile.encoding=UTF-8 Picked up JAVA_TOOL_OPTIONS: -Dfile.encoding=UTF-8 scan complete in 2ms Connecting to jdbc:hive2://analytics1003.eqiad.wmnet:10000 Connected to: Apache Hive (version 1.1.0-cdh5.5.2) Driver: Hive JDBC (version 1.1.0-cdh5.5.2) Transaction isolation: TRANSACTION_REPEATABLE_READ Beeline version 1.1.0-cdh5.5.2 by Apache Hive 0: jdbc:hive2://analytics1003.eqiad.wmnet:100>
Ctrl+C to exit.
beeline --help, this should show you all the available options.
We have a wrapper script setup around beeline that sets defaults for the database url, username (current user) and outputformat (tsv2). Any of these can be overridden by passing the option while invoking beeline.
Running queries works the same way as the hive CLI - you can read query from a file using -f, pass it as a string using -e, etc. See Analytics/Cluster/Hive#Querying.
Beeline usage, and all the different options are explained here - https://cwiki.apache.org/confluence/display/Hive/HiveServer2+Clients#HiveServer2Clients-Beeline–CommandLineShell.
Beeline has some cool features/bugs solved over the old Hive client.
- View results with huge number of columns in vertical alignment with --outputformat vertical