To issue queries against Kudu using Hive, one optional parameter can be provided by the Hive configuration:
Comma-separated list of all of the Kudu master addresses.
This value is only used for a given table if the kudu.master_addresses table property is not set.
For those familiar with Kudu, the master addresses configuration is the normal configuration value necessary to connect to Kudu. The easiest way to provide this value is by using the
-hiveconf option to the
Though it is a common practice to ingest the data into Kudu tables via tools like Apache NiFi or Apache Spark and query the data via Hive, data can also be inserted to the Kudu tables via Hive INSERT statements. It is important to note that when data is inserted a Kudu UPSERT operation is actually used to avoid primary key constraint issues. Making this more flexible is tracked via HIVE-22024. Additionally UPDATE and DELETE operations are not supported. Enabling that functionality is tracked via HIVE-22027.
INSERT INTO kudu_table SELECT * FROM other_table; INSERT INTO TABLE kudu_table VALUES (1, 'test 1', 1.1), (2, 'test 2', 2.2);