is it ok (as a good practice) in OBIEE Answers, when dealing with a complex BM, to build a big report with lots of dimensions and measures involved , and then create many views for that report, each using a limited number of dimensions / measures available in the report ?
I have experienced a few cases when a complex report with multiple views displays incorrect results , while if creating simpler reports for each of the views the results are correct.
and that seems to me because, on each of the small simple reports a correct sql query is generated, while on the big report OBIEE tries to achieve the requirements using just one or two much more complex (and with lower granularity) queries ( the result of which are are used by presentation layer to further compute -sometimes incorrectly- the values for each of the report views).
sometimes even when I exclude a dimension column from all views (while is still present in the report) the sql query generated still use it.
is this a normal behavior ? have anyone experienced similar issues ? I'm having some hard time explaining these to the users that create reports.
What you got is write and its normal behavior. If report is simple then we can keep couple of column in excluded (while is still present in the report) section but as you said complex report I wont suggest.
Couldn't find anything in the documentation to support that , though. How to explain this to the report writers ? many times it is more convenient to them to just create a big report with lots of views...instead of many reports.