Thank you for your reply.

My first concern here is to "view", but I could imagine the need to edit this special view.

Let's suppose I have a log file - not a server log, but a debug output log. This could be something like
---
Info: abc
Error: description1
Info: cba
Error: description2
Trace: aabbcc
Trace: dddd
Error: description3
---
much, much longer of course. What I would like to view is
---
Error: description1
Error: description2
Error: description3
---
i.e., hide lines that don't start with "Error:" (or, with other words, show only such lines). Such, I would have a very good overview of errors.
I cannot delete the lines between Error lines, because context will become important later on. I would like to find certain errors or certain details in the description and "expand" the file around this special error.
Maybe I'd like to add some comment to the lines for a colleague ("probably not important for the bug to be found" a.s.o.) - which is where the "edit" comes into play.

Do you think this use case it too uncommon? I, at least, could imagine similar requirements for csv files.

Maybe you could see it like a special kind of folding - fold everything between errors. In the meantime I remembered VIM, which has a folding by expression mode. Of course this is a bit complicated. It probably could be simplified for certain cases.