Keeping track of changes in a large data set?

[btrinh]

Hello! Hope someone could help me with this problem.

I have a large dataset (~2mil records) that is updated periodically, and I need to keep track of the changes in this set. Whenever there's a new update, I'll receive the new data dump with ALL records, not just the updated ones. It looks like this:

Update	Period	Item	Value
Update 1	2020-01	A	1000
	2020-01	B	500
	2020-01	C	4000
Update 2	2020-02	A	1000
	2020-02	B	1000
	2020-02	C	2000
Update 3	2020-03	A	1000
	2020-03	B	1000
	2020-03	C	3000
Update 4	2020-04	A	1000
	2020-04	B	2500
	2020-04	C	3000
	2020-04	D	500

My question is, what is the best way to model the data to keep track of Value changes?

If I keep all historical records like the above example, I'll end up with ~8 millions records after just 4 updates.

But if I keep only the changes, or the latest record + changes if any, I cannot visualize the changes in a pivot table (the final requirement). Before() doesn't work, since the data is missing:

Looking forward to your ideas! If you need any clarifications, please feel free to ask.