Once you've gotten to the point that you need "advanced" bash scripting, you're far beyond the point where you should have just rewritten your script in a real language, where simple common easy things aren't considered "advanced".
bash (or shell) has the advantage of being able to run nearly anywhere, which is why it is so ubiquitous. Other languages used for scripting that have more syntactic sugar like python usually come with a lot of bloat and headache with packaged libraries and dependencies that need to be installed.
shell scripts have external dependencies often but are typically resolved with a quick one liner within the script itself - such as installing jq if it can’t find it on the system.
I'm not a professional python programmer (nor an native English speaker) but for my tiny needs I resolved the library/modules import problem with pyinstaller which creates a single binary bundle from the virtual-env you invoke it. Just be sure the target machine has the same OS (or more recent): glibc and other sys libs are not bundled.
Bash runs almost anywhere but many of the tools you need for bash, external binaries, are not ubiquitous or have subtle differences between versions and OSes. But Python stdlib is always the Python stdlib.
It's just as easy to install Python as it is to install jq, and Python actually has its own json parser so you don't have to bend over backwards and eat shit to consume and produce something as standard and ubiquitous as json. And Python runs nearly everywhere too.
And it doesn't have to fork off a subprocess and serialize and deserialize text for every single little json structure or property you want to extract or combine.
Manipulating json from bash via jq may seem advanced high tech power user stuff to somebody who only knows bash programming and jq, and refuses to learn python or any other language, but believe me, manipulating json is extremely trivial and massively more efficient and expressive in a real scripting language.
Doing simple things in the most difficult complex way possible doesn't make you cool, it just makes you inefficient and ineffective. If there's any json manipulation or any other data structures than strings, bash is simply the wrong tool for the job, no matter how many of jq's command line parameters and its domain specific mini-languages you can remember.
It's not like there are any modern computers that it's not trivial to install Python or any other language on. If you're using an old antique MicroVax or Cray-1 or PiDP-10 or Analytical Engine or Jacquard Loom that doesn't run the latest version of Python, then you're just doing recreational retrocomputing, and not serious about actually developing modern software on modern hardware for real world use. I mean, that's cool and all, but your choice in how to spend your leisure retrocomputing time should not affect how actual programmers solve real world problems today.
tldp is notoriously full of bad advice, especially when it comes to shell scripting. From my time on #bash back when freenode was a thing, this was the superior resource: https://mywiki.wooledge.org/BashGuide
