TM1 Forum

Hi Guys:

When I import a dimension which called "dealer" from a csv file.

And under the dim "Dealer" it has two hierarchy, one hierarchary is from the view of Region. The other is from the view of group.

The first hierarchy:
Total dealer group
Dealer Group
Dealer Code

The senond hierarchy:
Total dealer Region
Dealer Region
Dealer Sub Region
Dealer Code

I want to realize this requirement so I define two TI process to run different hierarchy. The Total Dealer Region and the Total Dealer Group are on the same level

At first I run the first hierarchy is success. And then I run the second hierarchy is also success. But when I add some data to the files. Run the first process then get the warning.

Error: Epilog procedure line (13): The resulting hierarchy for the dimension "Dealer" does not align with the level structure defined in the mapping. When parent-child dimensions are being treated as level based or when level based dimensions are being constructed using source data containing non-unique member identifiers, unexpected hierarchy structures may result. For the case when non-unique member identifiers exist in the source data, turn on the "Qualify member names" dimension property in the mapping page.

How to make the warning disappear?

The process wizard seems to not like it because on the second run the dimension already contains MORE levels than you are trying to import for. I don't use the wizards so cant give much detail but I imagine having one csv file with both hierarchies will work for a single wixardefwizarded process... or my preference just write the TI or TIs manually and it will be fine.

declanr wrote:The process wizard seems to not like it because on the second run the dimension already contains MORE levels than you are trying to import for. I don't use the wizards so cant give much detail but I imagine having one csv file with both hierarchies will work for a single wixardefwizarded process... or my preference just write the TI or TIs manually and it will be fine.

I am a new on the TM1. You just mention "write the TI or TIs manually and it will be fine" what does it mean? I use the Performance Modeler to develop.

I am a new on the TM1. You just mention "write the TI or TIs manually and it will be fine" what does it mean? I use the Performance Modeler to develop.

I think what Declar meant was to write TI process on your own by declaring variables contents as 'other', Write your own TI code by defining dimension names,top nodes elements as well as attributes if you need any in prolog and metadata depending on your requirement.

I don't have any Performance Modeler experience so can't tell much about it,But If you want to go TI route ,Please also go through TI guide,It has pretty good example on page no 89 under the heading of Advance Scripting and Dimension functions in Reference guide ,Also Bedrock has quite few examples of TI.

Thanks BariAbdul.

BariAbdul wrote:

I am a new on the TM1. You just mention "write the TI or TIs manually and it will be fine" what does it mean? I use the Performance Modeler to develop.

I think what Declar meant was to write TI process on your own by declaring variables contents as 'other', Write your own TI code by defining dimension names,top nodes elements as well as attributes if you need any in prolog and metadata depending on your requirement.

I don't have any Performance Modeler experience so can't tell much about it,But If you want to go TI route ,Please also go through TI guide,It has pretty good example on page no 89 under the heading of Advance Scripting and Dimension functions in Reference guide ,Also Bedrock has quite few examples of TI.