o*********y
发帖数: 70 | 1 在OLAP中用的多吗，目前有独立的MDB vendor吗？这个方向发展前景如何，先谢了。 | e****7
发帖数: 4387 | 2 vender 方面 wiki 一下 olap.
工作市场还不错，有relational model 经验会有帮助。 | e****7
发帖数: 4387 | 3 vender 方面 wiki 一下 olap.
工作市场还不错，有relational model 经验会有帮助。 | c*****e
发帖数: 1106 | 4 Many BI tools would present a logical multi-dimensional view, but in terms
of products with a physical MOLAP engine. MS Analysis Service and Oracle
Hyperion Essbase are the two most prominent
ones. Essbase is a much more pure MOLAP. Both support the rather popular
query language MDX.
SAP BW is another one but it is pretty much tailored for its apps. Oracle
has another in database product call Oracle OLAP option, but it is suffering
from an obscure language.
Physical MOLAP can be a good solution for problems where there is a well
defined universe and analysis to have cubes prebuilt for the intended
purpose. However, when the universe becomes much bigger and/or analysis
becomes more ad-hoc, it become much hard for a physical MOLAP engine to
solve.
That's why in the last 10 years, there have been many new MPP (DB) products/
startups to handle increasing data volumn and ad-hoc analysis, but virtually
no new entries in MOLAP arena.
Having a physical MOLAP engine does have other advantages. Many products try
to present a logical multi-dim model, i.e. hierarchical view of relational
data; whichever query languages they use on the front-end, there is bound to
be loss in translation to sql both in terms of performance and express-
ability. |
|
