Trying to understand the User Defined Functions for integration between DB2(also Netezza) and Jaql.
The doc for HDFS_READ says that the HDFS files must be in CSV form(CSV for DB2, Netezza should be pipe delimited but allows specifying some other delimiter).
Just want to verify that these UDFs are implying that I need some process to get my HDFS files into that form (CSV for e.g.) before using the UDFs?
My files in HDFS are key/value pairs and not in CSV form. Do I need to convert to something like this: "key,value" for DB2 and "key|value" for Netezza?