I'm new to OLAP, cubes, star schemas...
I'm using Analytic Workspace Manager Version 10.2.0.1.0.
Is it possible to drill down without creating hierachies ?
Once hierarchies are created and i want to do the mapping, it seems that every level needs to have a key, am I wrong ?
Therefore, if I take a Time table that I have to map, I should have this king of physical structure :
Another example, if I have a Product table, I should have this kind of physical structure :
Thank you for your help.
The dimension members have to be unique in OLAP.
You can concatenate level id with the dimension member to make dimension members unique at each level of OLAP dimension. This is a common practice.
Lets say one row in your relational TIME dimension table is: year_id = 1 year_desc = '2011' quarter_id = 1 quarter_desc = 'Q1-2011' month_id = 1 month_desc = 'Jan-2011'
and you level ids in olap are defined as 'YR_LVL', 'QTR_LVL', MTH_LVL'
This can be loaded into your OLAP time dimension as 3 time dimension members: time = 'YR_LVL_1' time_long_description = '2011' time = 'QTR_LVL_1' time_long_description = 'Q1-2011' time = 'MTH_LVL_1' time_long_description = 'Jan-2011'
You can either do the concatenation setting in OLAP time dimension definition in AWM (without making any change in relational table or view)
or in your source sql view that is "mapped" to OLAP time dimension.
In other words, please correct me if i'm wrong, if the relational table doesn't have year_id, quarter_id, month_id as three different columns or doesn't have any data to define them, I will need to build a Time Dimension with a structure defined by levels and keys for each level ?
Thank you again