Integrating CHYT with external systems

This instruction demonstrates how you can integrate CHYT with external systems. Like ClickHouse, CHYT supports interfaces for interaction with external systems via HTTP/JDBC.

Connecting Trino to CHYT

1. Open ClickHouse ports in the http-proxy configuration.

2. Set up Database-as-Directory.
   Under YT config in the selected clique's speclet, specify the path to the directory that will be represented as a database in ClickHouse terms:

   {
       ...
       "database_directories":
       {
           "dbo": "//home/my_db_folder"
       }
   }
   

   dbo: Sample name of a database that will contain tables from //home/my_db_folder.

   Once this setting is specified, queries to the clique will look like this:

   SELECT * FROM dbo.table_name;
   

   This will be equivalent to:

   SELECT * FROM `//home/my_db_folder/table_name`
   

3. Set up Trino
   Trino connects to CHYT via a built-in ClickHouse connector. Create a connection configuration file /etc/catalog/ch.properties:

   connector.name=clickhouse
   connection-url=jdbc:clickhouse://my_proxy_url:8123/YT?chyt.clique_alias=my_clique_name
   connection-user=my_clique_name
   connection-password=my_token
   

   where my_proxy_url is a proxy URL, my_clique_name is a clique name, and my_token is a YTsaurus access token.

After that, the ch catalog will be available in Trino, with tables located on the YTsaurus cluster and accessible via the my_clique_name clique.