I've looked at plenty of perl and even maintained a little, but never fully grokked it, not implemented something from scratch in it.
Though I understand what you are saying, I sometimes feel the same with combinations of awk, sed, and other utils.
Though I suspect perl could be more powerful for that kind of processing.
Perl is so ubiquitous that makes awk and sed pretty redundant, if not outright worse. Perl regular expressions are so much better than whatever ugly mishmash sed uses that I haven't touched it in years.
I usually use perl with a combination of cut, grep and sort. If I had awk and sed in muscle memory, Iād definitely use them in some places where I use perl (awk and sed require fewer keystrokes for certain tasks).
I also like using perl for non-trivial stuff. For instance: parse a file, stick stuff in a hashtable or two, then iterate over the keys in sorted order, check a bunch of invariants, and then emit some other wonky format by chasing keys in the hashtables.
Some day, I hope to master the techniques in this book:
https://en.wikipedia.org/wiki/The_AWK_Programming_Language
which includes an RDBMS and assembler + interpreter, and then culminates in a 4 page section that implements make.
At that point, I might be able to switch away from perl.
I still use one-liners all the time, and when I need to implement a non-trivial, but small stream transformation, the resulting perl is << 1/10th the size of the unmaintainable python my coworkers would write.