I'm not sure which direction to take - has anyone dealt with GA4 weekly, monthly, and yearly datasets before?
I've realized that if I need to get a summary number of users over a certain time period in GA4, the only true way to make sure they are not double counted is to make separate datasets that pull in users by day, by week, by month, and by year.
If, though, I make a dataset that pulls in users per week and that appends and updates daily, it pulls in duplicate data - On Day 1 of the week, it pulls in Day 1 data. On Day 2 of the week, it pulls in users from Days 1 and 2. On Day 3 of the week, it pulls in users from Days 1, 2, and 3. So by the end of the week, my summary number for the week is massively inflated.
If I make it replace, the data will start deleting after 14 months, which is the longest time Google is storing that data.
So, for what I can figure, I need to either schedule the dataset to only update once a week, and for the previous week only, or I need to set up an ETL for the Weekly, Monthly, and Yearly datasets to only accept the last data update of any week, month, or year.
Am I missing something? Is there a recommended course of action for this?
