2. Because it's closer to an amalgamation of the standard shell scripting tools (cut, sort, jq, etc) than it is to a SQL variant.
so, it remains simple and concise for the easy problems, yet can scale up to address the trickier ones as well.
i've used csvkit, sqlite, jq, etc. miller is my favorite tool for data-fu (with sqlite being a close 2nd)
2. Because it's closer to an amalgamation of the standard shell scripting tools (cut, sort, jq, etc) than it is to a SQL variant.