-h or --help

This option prints the impala shell help details as console output. You can use this option as follows:

$ impala-shell –h or $impala-shell --h elp

-v or --version

These options print the current running impala-shell version and you can use this option as follows:

$impala-shell –v or $impala-shell –version

An example of using the –v option is as follows:

$ impala-shell -v
Impala Shell v1.1.1 (83d5868) built on Fri Aug 23 17:28:05 PDT 2013

-V or --verbose

This option enables the verbose output mode, so you will see lot more information in your output.

--quiet

This option disables the verbose output mode.

-c

This option continues command execution even when query failure occurs. This option is useful when multiple queries are submitted via shell scripts or other means, and processing does not stop if one or more queries return failure.

-I hostname or –impalad=hostname

This option connects impala-shell to impalad, using a specific host-passed hostname. By default, connection is made on port 21000; however, a different port number can be passed with the hostname as hostname:port. This command is useful when impala-shell is connected to such an impala daemon, which is running on a machine other than the current machine running impala-shell. You can use this command as follows:

$impala-shell –l myhostname:21005

-r or –refresh_after_connect

This option refreshes Impala metadata after successful connection. This option is the same as applying the REFRESH statement after connecting to the Impala server. You can use this command to refresh metadata just after connecting to impalad as follows:

$impalad –r 

-d database_name or –database database_name

By default, when impala-shell starts and connects to the Impala server, it connects to a database named default. Using this option, you can pass a desired database name to be used as the default after a successful connection. If you have a database named logs and want to use it with impala-shell, you will achieve your objective by using the following command:

$impala-shell –d logs

-q query or –query=query

This option allows you to issue a single query directly from a command line. This option is most useful when you have a need to run impala-shell inside a programming language such as Python or Perl, or inside a shell script.

If launched from the shell script, the query output will be displayed on the console or can be collected through the programming interface, if used. An example of using this option is as follows:

$ impala-shell –query ''select * from table_name;'

-o filename or --output_file filename

This option allows you to save the query output to a file.

-B or --delimited

This option is use to produce comma-separated, tab-separated, or other delimited text files as output. The default delimiter is the TAB character ( ); however, it can be changed using the –output_delimiter option and passing the desired delimiter character.

-f query_file_name or –query_file=query_file_name

Using this option you can pass a SQL query file that includes all of the queries you want to pass through impala-shell. An example of using this option is as follows:

$impala-shell –f myqueryfile.sql

-p or --show_profiles

This option is similar to the SQL EXPLAIN statement that provides a query-execution plan including detailed query execution steps for every query executed by the impala-shell

-k or --kerberos

When using this option, the impala-shell connects to impalad using the Kerberos authentication. Make sure to have the Kerberos authentication enabled and working with impalad, otherwise a connection cannot be established and errors will be displayed.

-s Kerberos_service_name or –kerberos_service_name=Kerberos_service_name

This option instructs impala-shell to connect to impalad using a particular impalad service principal, as passed in the option. If Kerberos_service_name is not specific, Impala is used by default with Kerberos.