Hi all,
I am getting closer to the magic formula of estimating cube size.
Mainly for future growth sizing consideration.
I encourage you to reply and leave feedback till we find more accurate one.
Partition weight=
( ( (4 * number of granularity attributes) + (4 * number of measures) ) * number of records ) / 3
* This assumes 1/3rd compression overall and an average of 4 byte measures (if you use long integers or double then you’d multiply measures by 8).
Have fun!
Ronen