I agree with @leocwlau. I would be tempted to track both the ID and the Name in the implementation (Name for readability, ID for uniqueness).
You might even want to pass both values into the same dimension with a delimiter, and use Classifications to split it to Name and ID:
Example:
PCDCR232452552 | Document Title Sample
I am using a space, pipe, space ( | ) as my delimiter. Then on that one dimension (which I would rename Document ID and Name), I can create 2 classifications (Document ID and Document Name), and use simple Regex rules to split the values:
Regex: ^([a-zA-Z0-9]*)\s|\s([\sa-zA-Z0-9]*)$
Which should take "PCDCR232452552" as group 1 ($1) and "Document Title Sample" as group 2 ($2)
So basically I would only need 2 rules.
Now I can show the original "ID and Name", or just the "ID" or just the "Name" depending on my needs.
But if you don't have the name today, then you can create either Classification Rules in the Rule Builder (if you have a smallish number of documents that you can map one to one, or you can use the Classification Importer to map each ID to a Name if you have a large number of documents.