Hi folks,
I'm on my latest kick which is experimenting with hierarchies and in doing so I've made some realizations. For starters, I noticed that for each column in my analytic view there already seems to be an auto generated hierarchy in the _SYS_BIC schema, one for each and every column. Conversely for attribute views there seems to be a single hierarchy generated regardless of how many columns exist in the view. The single column seems to be based upon the first KEY listed in the OUTPUT of the view. (Although this seems to be overwritten if I create my own custom hierarchy) Now I have a few questions based upon these discoveries;
1) Each auto generated hierarchy seems to contain a distinct list of values for that column. Is this why activating a particularly large view could take a long time, because each of the hierarchies are being populated?
2) Are these auto generated hierarchies somehow used for decompression when loading data from disk into memory?
3) When I experiment with unloading (FROM MEMORY) the source table for a particular column, I notice the distinct list in the hierarchy remains intact. Does this mean this distinct list is actually stored on physical disk or is it in memory? (With this one i'm just trying to understand how/when this distinct list is updated)
These questions are to satisfy my curiosity, not to solve an immediate business need. Thanks for any comments you may have on this.
-Patrick