Number of fed cells in }statsbycube

[rksharmatm1]

I was trying to understand overfeeding and started looking into “}statsbycube” cube. In detail, what the values “no.of. Fed cells” refers t etc etc..? This could be used to check overfeeding in cubes and then restricting rules in possible cases only to specific elements (instead of all elements ) of a dimension. Probably, this will reduce no.of fed cells/memory used by the feeders for that cube. Please add your comments or correct me if my understanding is not correct.

Cube 1 Dim1: 10 elements (Elem1, Elem2….)
Dim 2: 3 elements (s1, s2, s3)
Dim 3: 1 element (Measure)

#Rule in Cube 1 ( no other rule for this cube) :
[‘Measure’] = N: 100* DB (cube2,!dim1,!dim2,’Amount’);
#Feeder in Cube 2:
[‘Amount’] => DB (Cube1, !dim1,!dim2,’Measure’);

as we can see rule applies to all 10 elements of dim1, all 3 elements of dim2 in Cube 1. So, no.of. Fed cells should count to 10*3*1 = 30 cells
Q1: Is my understanding correct? (Or are there any other factors included while calculating total no.of. Fed cells for a cube. Specifically, I’m referring to "no.of. fed cells" value in “}statsbycube” cube)

####
Now, if we change the rule in Cube1 to restrict this rule only Elem1 of dim1 as below, and keeping the feeders in Cube 2 untouched.
[‘Elem1’,’Measure’] = N:100* DB (cube2,!dim1,!dim2,’Amount’);

Q2: will this reduce "no.of. fed cells" value in “}statsbycube” for Cube1 to 3 cells (instead of 30 cells) in ?

Thanks in advance for all your comments and help.!!
Best Regards,

[rmackenzie]

as we can see rule applies to all 10 elements of dim1, all 3 elements of dim2 in Cube 1. So, no.of. Fed cells should count to 10*3*1 = 30 cells
Q1: Is my understanding correct? (Or are there any other factors included while calculating total no.of. Fed cells for a cube. Specifically, I’m referring to "no.of. fed cells" value in “}statsbycube” cube)

No, just because the rule applies to 30 cells doesn't mean all 30 cells are fed. The number of fed cells will depend, in your example, on the number of non-zero cells in cube2 that are within the scope of the left-hand side of the feeder.

In fact, because any cell can feed any other cell then you could have feeders in cubes 3, 4 and 5 feeding cells in cube1 where those feeders have nothing to do with the rule applying to the 30 cells. This is a pretty important concept to get your head around.

Now, if we change the rule in Cube1 to restrict this rule only Elem1 of dim1 as below, and keeping the feeders in Cube 2 untouched.
[‘Elem1’,’Measure’] = N:100* DB (cube2,!dim1,!dim2,’Amount’);

Q2: will this reduce "no.of. fed cells" value in “}statsbycube” for Cube1 to 3 cells (instead of 30 cells) in ?

No, changing the rule won't change the number of fed cells. To change the number of fed cells you need to change the feeder in cube2. Also note that just changing the feeder might not 'un-feed' previously fed cells. You need to unload and reload the target cube to be sure that this has occurred.

Changing the rule just changes the number of cells that should be fed, by some feeder, in some cube; which for your example happens to be the feeder in cube2.

[rksharmatm1]

[quote="rmackenzie"]

Thanks a lot for your explanation. Things are more clearer to me now on how these figures are being calculated and referred to. Will try to perform some tests on sample cubes. Thanks.

Best Regards,
Ravi Sharma