Another Regression Bug :@ (SUBNM)
Posted: Wed Sep 03, 2008 11:42 am
I'm starting to get a little more than annoyed with basic functionality that is broken in TM1...
I've tried this on both the dev and prod servers that are using slightly different versions of 9.0 (dev is 9.0.3.182 update 7 / prod is 9.0.3.173) and both do the same thing. Its critical to one of my procedures to remove errors from data before it's submitted to us, and it simply doesn't work.
I have a subset in a raft of dimensions called "Validation List". It's built using MDX to grab only level zero elements AND only when they do not have a value in an attribute called "ValidationList". It allows us to exclude certain elements which are valid but we don't want people using, as well as exclude the usual mistakes of entering consolidations insteda of nodals. There is NOTHING wrong with these subsets, they are spot on. The problem is when SUBNM is used against them.
SUBNM can grab values from a subset via either index OR element name. If i write SUBNM using the index, it's fine - all i get are the elements that existing within my subset. But, if you do the same using the element name, its completely ignoring the subset and looking at the entire dim. That means my SUBNM returns a value even when checked against a consolidation or worse still, one of the "banned" elements.
I'm relying upon this to allow users to vet their data (as they are the only ones that know the valid entries, not us), and in Excel its the only way i can do it without going back to the old method of manually creating a validation file that each data submission is linked to.
Does anyone else using a version of 9.0 manage to get SUBNM to work correctly i this manner, and if so which version is it with? I'm desperate for a solution today, and i'll be ringing Cognos Support right after lunch, not that i hold out much hope as they still haven't given me an answer on the screwed up ViewCreate process in TI...
I've tried this on both the dev and prod servers that are using slightly different versions of 9.0 (dev is 9.0.3.182 update 7 / prod is 9.0.3.173) and both do the same thing. Its critical to one of my procedures to remove errors from data before it's submitted to us, and it simply doesn't work.
I have a subset in a raft of dimensions called "Validation List". It's built using MDX to grab only level zero elements AND only when they do not have a value in an attribute called "ValidationList". It allows us to exclude certain elements which are valid but we don't want people using, as well as exclude the usual mistakes of entering consolidations insteda of nodals. There is NOTHING wrong with these subsets, they are spot on. The problem is when SUBNM is used against them.
SUBNM can grab values from a subset via either index OR element name. If i write SUBNM using the index, it's fine - all i get are the elements that existing within my subset. But, if you do the same using the element name, its completely ignoring the subset and looking at the entire dim. That means my SUBNM returns a value even when checked against a consolidation or worse still, one of the "banned" elements.
I'm relying upon this to allow users to vet their data (as they are the only ones that know the valid entries, not us), and in Excel its the only way i can do it without going back to the old method of manually creating a validation file that each data submission is linked to.
Does anyone else using a version of 9.0 manage to get SUBNM to work correctly i this manner, and if so which version is it with? I'm desperate for a solution today, and i'll be ringing Cognos Support right after lunch, not that i hold out much hope as they still haven't given me an answer on the screwed up ViewCreate process in TI...