Since Hive v2.0.0, the Hadoop Procedure Language SQL (HPL/SQL) (http://www.hplsql.org/) available to provide store procedure programming in Hive. HPL/SQL supports Hive, Spark SQL, and Impala, and is compatible with Oracle, DB2, MySQL, and TSQL standard. One of its benefits is making the migration of existing database-stored procedures to Hive easy and efficient. Using HPL/SQL does not require Java skills to implement what can be done through UDF mentioned. Compared with UDF, HPL/SQL's performance is a little slower and it is still new for production usage.
The following is an example of creating a stored procedure. HPL/SQL supports the creation of both Function and Procedure:
$ cat getEmpCnt.pl
CREATE PROCEDURE getCount()
BEGIN
DECLARE cnt INT = 0;
SELECT COUNT(*) INTO cnt FROM employee;
PRINT 'Users cnt: ' || cnt;
END;
call getCount(); -- Call a procedure
In order...
