Using Content Importer (Tridion Docs 14 SP3), we are stumped for the moment on why the value of href in map/topicref is being forced to uppercase during import?

We have not ruled out user error in the preparation of maps for import.

I know there are parameters in XML Write Plugin Settings that will do this on a Create or Check-in action. Does Content Importer access these? Or, is there a setting in client appdata for Content Importer? Or, is there a "secret" XSL file delivered by RWS that is doing this?

The problem is that the value of href in the file to be imported is not "just" an ID; it's the full export name you get when doing a DITA XML publish from Tridion Docs. 

So at some point in the process, something like <topicref href="Title=GUID=Version=Lang=[Res].dita>, everything in href is getting uppercased, not just the GUID ID piece of the delimited file name. This breaks the link to the object name in TD14 which actually retains its upper-lower case quality. Especially, I suppose, the Lang piece of the file name (that is, "EN" does not equal "en").

This also begs the question, is the value of href shown in the screenshots even valid? Is it allowed to use the full, delimited file name in the href, or should it be only the GUID ID before running Content Importer on it? Is it intended that Content Importer  will resolve these? Because it seems they are not being resolved (title/lang/ver/resolution pieces not being removed from href value).

Or, are we simply using Content Importer incorrectly?

Before import:

After import, as viewed in TD14: