Margin[dynamic]->West[dynamic]: according to the dynamic calculation order:
sparse->account->time->dense->two pass
firstly, the virtual block [west] will be created in memory, then margin will
be calcuated dynamically in this block based on sales and COGS
now consider the different scenario with combinations of batch/dynamic
calculation, dense/sparse dimension:
Margin[dynamic]->West: West has to be aggregated before see this
Margin->West[dynamic]: Margin has to be calculated before view this
Margin[dynamic]->Q1: Q1 need to be aggregated firstly
Margin->Q1[dynamic]: Margin need to be calculated firstly
100[dynamic]->West: West need to aggregated firstly
100->West[dynamic]: 100 need to aggregated firstly
so basically non-dynamic member need to be calculated firstly before
dynamic calculation, no matter the non-dynamic member is in dense or
sparse dimension.
as far as calculation orders inside dense dimensions (including Account and
TIme) and sparse dimension, it has no difference between batch calculation
and dynamic calculation:
dense: account->time->dense dimensions along outline
sparse: sparse dimension along outline
why dynamic calucation for sparse dimension member could have
performance issue? say West is dynamic, everytime read West, essbase got to
read blocks for California,Oregon,Washington, Utah, Neveda 5 blocks! for
var analysis of actual and forecast, it have to read actual and
forecast blocks firstly.
for dynamic calcuation for dense member, it's in the same block
来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/8583032/viewspace-715886/,如需转载,请注明出处,否则将追究法律责任。
转载于:http://blog.itpub.net/8583032/viewspace-715886/
动态计算与性能挑战
本文探讨了在Essbase环境中动态与批处理计算的不同场景及顺序,特别是在稀疏维度下的动态计算可能带来的性能问题。文中详细解释了不同成员类型在动态计算中的处理方式,并对比了密集与稀疏维度下计算顺序的异同。
713

被折叠的 条评论
为什么被折叠?



