The point of it
It can be convenient to copy all the text files of one given month into one single text. It uses the date of each text. This function reduced 120 thousand single articles between 2005 and 2018 into 168 monthly texts, easier to handle when seeking trends.
How to do it
Choose a source folder, the file-types and a suitable results folder, and choose the delicacy required: years, months, weeks, days are the obvious choices.
Results
As each text file is found it gets added to the collection for the appropriate year/month/week/day in a folder of that name. This is as partial view of what I got:
Each month text file is correctly dated and is named using a YYYY_MM_(N).txt template, where N is the number of component articles inside.