I think the relevant point is not that if a dimension member value is empty
or not.
Remember the power of olap is Multidimensional analysis, so NON EMPTY is
usefull to unhide NULL CROSSJOINS, ie, if you have two dimensions that for a
certain level they do not have a valid cross (or for the hole tuple) then
NON EMPTY operator will exlude that tuple from the display.
Hope it helps,
Michael Prendergast
DataMat S.A.
Chile
"Elyo Ravuna" <msdn (AT) nebim (DOT) com.tr> escribió en el mensaje
news:F1F91FC1-952A-4255-B367-B513072F1D39 (AT) microsoft (DOT) com...
Quote:
According to KB#244650, null values in the fact table are treated as zero.
Therefore, there isn't any case where a physical measure is empty and
|
another physical measure in the same tuple is not empty.
Quote:
NON_EMPTY_BEHAVIOR accepts only physical measures. Since either all
physical measures in a tuple are empty, or they are all nonempty, it should
|
not matter which measure is selected as NON_EMPTY_BEHAVIOR.
Quote:
Am I correct, or am I missing something?
I would be very glad to hear a response from Microsoft.
Sincerely.
Elyo |