Overview

This section contains information about the YTsaurus command line: description, installation methods, and use cases.

Description

The YTsaurus CLI (Command Line Interface) is the most convenient way to interact with the YTsaurus system from the console. The CLI fully implements the features of all the commands supported by YTsaurus. The name of the command is specified as the first argument followed by its options. To get help on the CLI, just run the command:

yt --help
yt read --help

To work with a cluster, you need to get an access token and put it in the ~/.yt/token file or the YT_TOKEN environment variable.

For more information about the tokens and authentication, see the Authentication section.

In addition to the access token, you need to specify the cluster to work with. To do this, you can specify the proxy server name in the YT_PROXY environment variable: export YT_PROXY=<cluster-name> or pass the proxy server name to each command individually via the proxy option, such as --proxy <cluster-name>.

Below are some features of the CLI:

  • All underscores in command and option names were replaced by hyphens.

  • input_table_paths and output_table_paths are specified via the --src and --dst options.

  • To specify multiple columns for sort or reduce, pass the sort-by or reduce-by option several times, for example, --sort-by a --sort-by b will sort by composite key a b.

    Attention!

    If you write --sort-by a,b, sorting by one column named a,b will be performed.

  • The write command writes data chunk-by-chunk and under the transaction.

  • There is a find command (implemented on top of get) which partially replicates the unix find utility.

  • By default, the list command (without specifying the --format parameter) just enters a list of node elements, each on a separate string. It also has the -l option that will enter useful information about the nodes along with their names.

  • By default, any files uploaded into YTsaurus are uploaded without the executable flag. If you want to upload a binary file that will run in a job, you must specify the --executable option when uploading or set the /@executable=true attribute after uploading.

  • The formats of the data received at the command input and appearing at the command output can be specified through the following environment variables:

    • YT_ARGUMENTS_FORMAT: Manages the format for the values of the options of the --spec, --attributes type. The default value is yson.
    • YT_STRUCTURED_DATA_FORMAT: Manages the format of the get/set command. The default value is yson.
    • YT_TABULAR_DATA_FORMAT: Manages the format of reading/writing tables, as well as starting operation jobs. There is no default value, the variable value must be set explicitly or the --format option must be specified.

  • There is a method to create aliases for commands. Aliases are specified in the ~/.yt/aliases file in name=opts format. Example:

    mkdir=create map_node
ls=list
rm=remove
mv=move
cp=copy

  • Auto-completion of binary file commands and paths for bash is supported.