Problem with attribute distribution

hello,

i'm new ssas , have run issue way measure distributed among attribute.  essentially, have fact table of attendance records contain studentid, year , schoolid keys.  there absense type code , value of how of day student missed (0, .5 or 1) on day.  

now, apologize if can't describe issue propery, here best shot.  the individual entries in fact table (studentid, year , schoolid) seem rolling specific attendance dates in manner can't figure out nor want.

a query on underlying data such as

select attendcode, attenddate, sum(absense) attend_oos_detail  studentid = 'xxxxxx' 

group attendcode, attenddate

order attendcode

will return 46 distinct rows data distributed expect, when view deployed cube, end 46 days rolled 4 distinct days.  worse, of type 'o' codes rolled type 'a' codes (i've limited 2 simplify debugging process).  

looking @ detail of records ordered date, totals cube can reached summing between dates, regardless of attendance codes contained within.  for example range september 7, 2005 - april 18, 2006 listed  september 7, 2005 , contains sum of attendance entire range instead of being listed 1 day , attendance code.  

i'm omitting data for brevity, can include if helpful.

i've been playing various combinations of logical keys , whatnot, no end.  any appreciated.

thanks!

could describe dimensions in cube - have own separate dimension tables, or based on fact table? , how fact table related date dimension - granularity attribute?

 

SQL Server  >  SQL Server Analysis Services