Having 2 fields is definitely unacceptable for the current scenario. How about pushing the 400 characters limit?

Can you please explain the current scenario in more detail? I'm hoping we can find an easier solution for you that avoids the need to create such complex constraints. 

Hello Tessa,

Can you please tell me what information is unclear/missing in my request ?

We would like to know why your fields are configured this way, what business problem is being solved.

Document piles - composed of different document types, e.g. invoice, bill of lading, etc - are scanned; each page is OCRed and pattern matched to identify which document type it belongs to. So let say a pile has 500 pages; a workflow pattern matches each page as per its document type and ID, e.g. invoice #12345. That information is then stored in a multi-value field (one value per page). Afterwards, the user has to check if each page's information is okay and, if so, another workflow splits the pile into different Laserfiche documents based on consecutive pages, e.g. invoice #12345 runs from pages 9 to page 12, and so on. This issue is: if the workflow fails to pattern match a given page (or do so inappropriately), the user has to enter manually the information into the multi-value field. In doing so, I don't want to restrict the user in typing only lowercase or uppercase letters. For instance, 'Invoice 12345' and 'INVOICE 12345' need to be ok, not 'Invoice ABCDE'. So the constraint looks like ^[Ii][vV][oO][iI][cC][eE]\x20[1-9]\d*$. Given there are many document types, the multi-value field constraint exceeds the actual 400 characters limit.

Hope that helps ;-)

I see. Have you considered using Quick Fields for this scenario? This seems like exactly the kind of situation that Quick Fields was made for :)

Quick Fields would be able to run OCR (if needed), identify document types and split the pile into individual documents based on the pattern matching, and much more. It also has many tools to significantly speed up manual review so you could quickly fix any issues with the documents, including misidentified pages and OCR issues, as soon as they are scanned. 

This would avoid the constraint issues altogether, since there would be no need to record the problems in fields corresponding to each page. The problems could likely instead be fixed in-place by using Quick Fields.