[itsme]

It is how it is for good reasons and I think the current behaviour should not be changed just to allow a pretty special workflow.

I investigated that already and I'm still using such a trick to have files that are not songs sorted under # in the alfabetical list. The sort sequence of first characters that I use is: blank, +, ~
Many of the CSV files that I shared in the forum have two additional entry lines where title starts with ~, one for the complete book and one for the table of content (I usually CSV-import not complete books but just a few songs that I plan to use + these two)

In many cases SORT_TITLE, as mentioned by Mike, is probably the better and more consistent way to go. For maintaining purposes you can make it additionally visible in MSPs list views by adding it to "Title Format" in "Settings - Library Settings - Song Title Format". Here's an example:
%TITLE% %KEYS:- ${VALUE}% %SORT_TITLE:<${VALUE}>%
I backup some title formats that I use regularly in a text file and exchange them via copy/paste

Additionally (or alternatively) it's a good idea to use groups like collections or source types to simplify filtering. E.g. I have collections for separators and tables of content and I have a user filter for "only real songs"

You should be aware that special characters are handled and sorted differently in different contexts: file managers under Windows, Linux or Android, XLS files, text editors. In some cases some characters are even unusable or forbidden. The attached SortTest.zip contains a number of files that I used to play around.