The Kylin connector allows querying tables in an external Kylin database. This can be used to join data between different systems like Kylin and Hive, or between two different Kylin instances.
To configure the Kylin connector, create a catalog properties file in
etc/catalog named, for example,
Kylin.properties, to mount the Kylin connector as the
Kylin catalog. Create the file with the following
contents, replacing the connection properties as appropriate for your setup:
connector.name=kylin connection-url=jdbc:kylin://example.net/project connection-user=root connection-password=secret connector-planoptimizer-rule-blacklist=io.prestosql.sql.planner.iterative.rule.SingleDistinctAggregationToGroupBy
The connector-planoptimizer-rule-blacklist attribute is specially configured for kylin, and the default value is io.prestosql.sql.planner.iterative.rule.SingleDistinctAggregationToGroupBy
Multiple Kylin Servers
You can have as many catalogs as you need, so if you have additional Kylin servers, simply add another properties file to
etc/catalog with a different name (making sure it ends in
.properties). For example, if
you name the property file
sales.properties, openLooKeng will create a catalog named
sales using the configured connector.
you can access the
clicks table in the
SELECT * FROM Kylin.web.clicks;
If you used a different name for your catalog properties file, use that catalog name instead of
Kylin in the above examples.
Kylin Connector Limitations
The following SQL statements are not yet supported: