I've been confronted with the problem, that the case insensitive search command search, differentiates between ö / Ö, ä / Ä and ü / Ü. My question now is, how does splunk parse the german Umlauts exactly and will it support case insensitivity searches for these characters in the near future?
I'm aware of the workaround using eval lower() / upper(), so this question is not about solving the issue, but trying to understand where the issue really originates from.
There is an open issue being tracked for this problem (that also affects other non-ASCII characters), but there is no current fix version due to backwards compatibility concerns. I suggest you open a support case to add your voice to getting this addressed.
Sorry, I don't have a more promising answer for you at this point. Äusserst unbefriedigend... 😉