Your calculation is similar to
1000 * 1.1 ** (row_number() over (order by Date)-1)
min(Amount) over () * 1.1 ** (row_number() over (order by Date)-1)
first_value(Amount) over (order by Date) * 1.1 ** (row_number() over (order by Date)-1)
If the percentage is not a constant, will be a join from a table dynamically using a date column, how do we do it then?
I believe we need to use LN() log function also for this, please share your comments with any possible new equation.
2016-11-02, 1100 (10% of 1000 + 1000) + previous value
2016-11-03, 1232 (12% of 1100 + 1100) + previous value
2016-11-04, 1416.8 (15% of 1232 + 1232) + previous value
Correct, if it's a different constant for each row you need something like this:
SUM(Amount ) OVER (ORDER BY Date ROWS UNBOUNDED PRECEDING) * EXP(SUM (LN (joined_percentage_column)) OVER (ORDER BY Date ROWS UNBOUNDED PRECEDING))