Skip to main content

Hello, I’m trying to import a csv file (Employee referential) into a Dimension List (Employee Referential List). The target Employee List has 5 properties : ID (“generate unique ID” setting toggled), Name, Country, Cost Centre & Salary band. I’ve set the ID property with the “generate unique ID” parameter (+Integer format), because the CSV file has 2 employees with the exact same name (but not the same country).

I get an error message preventing me to run this process (in the “map data” step): “The unique Property “ID” has to be mapped”. (the source is grayed since it’s an automatic property.) Is there a problem somewhere? Pigment doesn’t allow a direct import to a dimension list having an “auto-generated ID” property? (Sorry if this was mentionned in one of the academy paths, but didn’t find this precise information. I just found that it is possible with a transaction list, but a dimension list should be more appropriate since we are talking about a referential update).

Below the main screenshots:

1/ The target dimension list

 

2/ The CSV Source file , to update the employee referential (target dimension list)

 

3/ When I want to run the import, the message “The unique Property ID has to be mapped” appears.

 

Thanks in advance (newbie in Pigment)

 

Hi ​@Raphael C,

I believe it is not possible to add items to a dimension list without mapping a column to at least one unique property that allows Pigment to identify whether it can update an item or has to create a new one.

Without mapping your data to a unique property, Pigment has no other option than treating all records as new, giving it a new unique identifier, and increase the dimension size with each import, even if there are no new employees. I would assume this is why it just doesn’t work without mapping an import column to a unique property.

The auto-generating ID is suitable for dimensions where the items are added manually by users through the “add item” action, it works really well.

Maybe someone else has a community reference or other supporting information to that, but I would say your best shot is getting some unique identifier into the file you want to import, either manually or in your source system. Ideally an HRIS employee ID, alternatively maybe a concatenation of employee name and company, or whatever your data has and is most feasible in this case.


Hi ​@Raphael C,

I believe it is not possible to add items to a dimension list without mapping a column to at least one unique property that allows Pigment to identify whether it can update an item or has to create a new one.

Without mapping your data to a unique property, Pigment has no other option than treating all records as new, giving it a new unique identifier, and increase the dimension size with each import, even if there are no new employees. I would assume this is why it just doesn’t work without mapping an import column to a unique property.

The auto-generating ID is suitable for dimensions where the items are added manually by users through the “add item” action, it works really well.

Maybe someone else has a community reference or other supporting information to that, but I would say your best shot is getting some unique identifier into the file you want to import, either manually or in your source system. Ideally an HRIS employee ID, alternatively maybe a concatenation of employee name and company, or whatever your data has and is most feasible in this case.

 

Hi Oliver, thanks for the detailed answer. Indeed, everything you said makes sense. I was just focused on a “first import/ Referential initialization” context, where there are no updates but just first creations of all occurences (where the auto-generated unique ID could have be used as the missing unique IDs you mentionned) . So more in laboratory mode (not a real use case). Thanks for your input.


Might be your best shot would be going through a transaction list, load the data from your file, then have a calculated property that gets the unique id generated and use it to generate your people dimension list ids.

I dont know if it would work that way. It looks like a bit of a convoluted way to achieve what you need, but i dont see any other option (if unique id column in the file is nlt feasible)


Reply