once process the records, make a copy of file and when new data come in to original file get the count both files and process the rows based on the count difference of original file & new data file.

repeat this every time

Hi @Naveen Kumar​ 

As you are describing, add one trigger to handle this situation for Every modification of the particular file, so that whenever your File got modified you will be able to run your code to update the record.

Simultaneously you have to store one Unique key value from your Present record so that you can get a query whether it appears New or not and saves your time by avoiding older records.

please let us know if you achieved your Goal.