To use the reference udf:

1. Create a folder containing the jar file in the unix environment;

2. ADD JAR path/reference.jar;

3. ADD FILE reference_file.txt;

4. LIST JARS; //verify the jar

5. LIST FILES; //verify the reference file

6. CREATE TEMPORARY FUNCTION reference AS 'pkg.UDFReference';

Then use the udf inside hive passing the file added to the hadoop distributed cache as second argument.