When missing code / unique identifier in transactual data (general ledger) for list dimension
What is the best way of treating a general ledger, in other words a big data transaction file, that is required to be in the data hub and is missing a unique identifier for each row, i.e. something to be used as a code for the list that will act as the dimension for the DATA module? I could create my own code in the file 1, 2, 3, ... and so on, but when the customer then load in the next month and they cannot start over from 1 but need to follow the count. This seems a bit manual. What would be the best practice here in case the customer will not / cannot provide me with a unique identifier?
Assuming the load is going to be manual:
- can you define a set a properties that would make up an ID (account, CC + ...) ? if yes you can use a property base import mapping
- prefix the count with the period. That means if you want to reload a month you have to clear the transactions
- have a list like 500K (any size as long as it is bigger than your load) items over time, and load it this way. Hence every month you can have the same ID0
I would say the unique identifier has to come from the source. Must read
Data Hubs: Purpose and Peak Performance - Anaplan Community
I agree with the earlier comments, the best way to handle this instance would be to create a unique ID using fields in your data source. The good news is you don't need to do this offline before import, you can instruct Anaplan to combine these fields in your import process0
While this does work, it is best to create a unique ID at the source to remove extra processing time and work by Hyperblock. Also, by creating the unique member at the source, you will be able to import the transactional data to Trans module and the properties to a SYS Properties module. If you use the combination of properties method, you will not be able to do that and will be forced to store all data as properties in the list which is not advisable and is against best practices.