Glad to hear that works for you! Though we can split the raw csv or json content into a smaller size just in case of OOM issue when doing the shifting stuff on JoltTransformJSON, it may cause other issue like: - result from JoltTransformJSON is not complete. That means, some part of the same user data ( with different bill_date)  may be wrapped in the other flow files. In such case, we will need to merge them back as a whole. - It would be better if we can keep raw csv data in some DB table. Then drag out the data rows with a specified limit--> we can use "split pages" way to query on such table,  - We can then easily fetch a logical completed data to do the rest shift things. With such completed result, we don't need to worry about the data is not an info-incomplete one. ... View more