Can a wildcard be used for document searches with the Unity API if the keyword has a mask value?

[Grant_Fish]

I am trying to use the Unity API document query to do a wildcard search for documents based on a keyword type that has a mask value. A simple example would be a keyword type that has a mask value of "0000A." This would allow values of 1234A, 1234B, 1234C, etc.

For a non-masked keyword, the normal solution is to use as asterisk ("*") at the end of the value and the Equals operator. So, keyword Last Name with value of "SMI*" will find documents with Last Name keyword beginning with "SMI." "Smith," "Smithson," etc. are returned.

If I try to use this same wildcard method to search for documents by a masked keyword, OnBase appears to reject the asterisk because it doesn't meet the mask requirements and an error is thrown. Referring to the example above, if I want to find all documents with a keyword value beginning with "1234" (1234A through 1234Z), I can't use "1234*" as the search criteria.

[Brian_Koning · Accepted solution]

Hello Grant,

Are you able to perform the same wildcard keyword search in one of the clients? I have a KeywordType configured with a mask of “0000A” and both the Thick and Unity Client interfaces prevent me from typing in an asterisk and performing the wildcard search.

I am able to type the asterisk when the mask is configured as “00009” which makes me think wildcard keyword searches and keyword masks do work, as long as the mask allows for the wildcard character.

[Grant_Fish]

This wasn't really the answer that I was hoping for, but if that is a limitation of the system then we will have to accept it, but I would like to see it changed. 

Whether there is a mask or not, the keyword value is just a string, and in certain situations searching for a partial match on a string is useful.

To give a very specific example, our application is used by child support enforcement agencies. Case numbers are assigned a numeric value for the non-custodial parent (the parent paying child support), and an alpha suffix to indicate which child. This is where my simplified example of 1234A, 1234B, etc. came from.

The correct mask for this case number format would be 0000A, since the last character must be A-Z. It would obviously be useful to search for all documents related to a particular non-custodial parent, for all children.

The existence of a mask, no matter how the mask is configured, should not prevent partial-match searching. Is there a possibility that this could be changed in the future?

[Brian_Koning]

I am not in a position to reliably comment on if this behavior can be changed or not. However, I suspect it will be very difficult to change given the functionality is central to OnBase and both the Thick and Unity Clients would need to be updated.

I have created SCR: #239686 for the Unity Api to track your request. It will be reviewed for inclusion with OnBase 17 or later. Feel free to reach out to your first line of support with any questions about the SCR itself.