I have a question related to import transactional data (Net Sales) into Data Hub, with the following requirements:
The Net Sales and Qty Volumes are grouped by more than 10 attributes required for Sales Planning Purposes. Best practices in Anaplan is to concatenate those attributes that will be the key for the line items, but in this case are more than 10 attributes so I think is not a good choice.
For this transactional it will be stored 3 years of data and it will be uploaded incrementally, means each month will be uploaded a delta.
Main concerns are:
What should be the key for the line items? Concatenation of more than 10 attributes or a sequential number (taking into account that import is done based on deltas)?
It should be the time a dimension or line item in the module?
Hi @jagalindo load the data to the stagging list(Numbered list) using combination of properties. Create the module using stagging list as dimension and in the module create a unique key(based on dimensions in data) for create the base list. Now use the base list to store the data.
Regarding the time being used as dimension or line item, better use it as line item itself if there are not many data records varing by time for unique key dimension that was created.
Please don't use combination of properties as it adds unnecessary overhead to the process. The list can be numbered or not, that doesn't matter because you shouldn't be using properties on that list and using combination of properties requires you do use properties. Please read this article which explains exactly why and how you should load your data.
The concatenation of ten properties is fine, in fact that is what you are supposed to do. Just be careful to use the codes of those attributes/properties because you have limited space available (60 characters). This will be your transactional list. Now, create two modules:
Transactional data which will be dimensionalized by the transactional list and Time. This is where your transactional data (Net Sales) gets loaded to.
The second module will be your Attribute or Properties module where you will, by using formulas, decipher the attributes from the code.
That will get your data loaded to the hub correctly. Now, since your are doing an incremental load from the source, which is PERFECT, now you need to figure out which records/data have been loaded so you only load that data (the data that has changed) to your spoke model. Please watch this video.