Description - As it stands, Adobe Data Feeds dropped their data as zipped TSV files. Header information is located in a separate location and must be joined to Hit Data files in order to get the column to header mapping.
Why is this feature important to you - In the space of "big-data", Parquet is becoming universal format. Additionally, it seems unnecessary to have the "schema" abstracted from hit data files.
How would you like the feature to work - In the option when creating a new data feed in the UI, Parquet should be an option. Even JSON format is acceptable.
Current Behaviour - Only TSV is offered and columnar information is stored in a different file