Thanks for the detailed explanation.
I have three more questions
1) I analyzed JSON file of downloaded dataflow. this JSON file has every thing like input & temp table details, join type and conditions, transformations, UDFs, and so on. Can I modify tranformation logic, input and output details in JSON file and upload back to Xcalar and process data?
2) can we use same dataflow to process different files ( Ex: file1, file2 and file3 contains 10, 7 and 50 columns respectively)
3) How to retrieve/download schema (column names, data type, length & scale) of final output if required?