Running into an issue with a particular dataset.  Trying to pull in Tower and Sub-Tower, but on some of the rows I get the dreaded {Various}.  I can't use LookupEx because it changes my total and row count.  I've tried adding things to the logic in addition to the Lookup to try and make it more unique, but that hasn't worked.  Is there maybe a formula for Row Label?  If yes, wondering if I could pull something like that in to make it unique?  In the meantime, will continue to try different combinations to get 'more unique', but if anyone has any ideas, please let me know - and thank you!  

Hi Jenny,

Can you share your Lookup formula? And what does your dataset look like in the column(s) you've tried using for the match?

FIXED, but leaving in case it helps someone else..

@John Adipietro would you believe I'm having this issue again?  Problem is, it's four years later and I'm at a different company, so can't go back to see what I did. 

I have two data sets - one with hypervisors and one with virtual machines.  I'm able to do a LookupEx to pull back hypervisors for all virtuals based on cluster name in each table.  I can't seem to get to one hypervisor per virtual, though, because there are multiple hypervisors per virtual.  Doesn't make sense to me, but I'm new and not on that side of the fence, so unsure how things are typically done. 

Either way, I need to get to one hypervisor per vm so I don't have various in there.  I thought about making the hypervisor name unique, but then there's still multiple hypervisors per vm, so if I pull back hypervisor based on vm, it won't do it - the dreaded various again. 

Is there a way to count unique rows?  So if hypervisor A is listed five times, then it'll show a unique number per row?  I've tried UniqueCount and RowCount and I'm not getting the results I thought I would. 

I looked at SBahuguna's comment below, but I'm not sure how to do that and looks like she's no longer with Apptio. 

Any ideas on how to whittle this down to one hypervisor per vm?  Have tried grouping it down as well, but that creates various on the vm side. 

Was leaning toward the unique count per hypervisor, and then filtering out all that are > 1, but I can't seem to figure out how to get the count added per unique hypervisor.

That's just me spit-balling, though - you probably have a much better way!

Just ran across this (Incrementing count with grouping ) and going to give that whirl..

Yeah that's not going to work because I can't create a transform per group of 100+ hypervisors.

I swear it feels like Lookup changed from R11 to R12 or I may be crazy.  Probably the latter.  But it seems like back with R11, Lookup would pull back the first result, but now it seems like it wants to pull all results and give me a various.  Back to the drawing board - again, and again, and again..

Added a Flatten Hierarchy step and that fixed it!!  I'm not using it how it's supposed to be used based on the link below, but it worked in my particular use case (so far).... Flatten dataset with duplicate children 

Hi Jenny,

Is it possible that you could do LookupEx and then fix the totals by dividing by the number of unique row counts? If not, as John mentioned, can we please have a look at the lookup column values and other data columns.

Sorry, I should have responded earlier.  I was fiddling with it over the weekend - even added the additional data to the unique identifier of the backing datasets, but it still would throw various.  I tried the LookupEx, but it was pulling in extra rows and increasing my overall amount (cost source master data).  I was trying to pull in tower and sub from another data source.  However, once I whittled down the dataset to only the rows I needed - one specific tower - it went away finally.   I'm hoping I'm set, but if it rears its ugly head again, it'll be back to the drawing board and I'll hop back on with additional details.  For now, I think I'm in the clear.  Thank you both and apologies for not closing this out sooner!

Not a problem Jenny. Glad you could resolve it. Please feel free to post it if you run into it again.

Thanks for this post!

@Jenny Franklin