Yeah I agree. In Oil I'm hoping to have the best of both worlds. That is, no silly quoting and parsing of strings, but also being able to use plain tools like "grep", and probably SQL.

https://github.com/oilshell/oil/wiki/Structured-Data-in-Oil

Note that R has some popular packages that allow SQL syntax over data frames.

So basically the history file could be a textual table. You could trivially import it into a database, or you can just grep it.

I see your "decorated" sql table, and I raise you my text file with five columns, that you can grep and sed and awk as if there was no tomorrow.

     cat history.txt | awk '$4 != 0'

"As if there was no tomorrow" - exactly, because once you add another column suddenly $4 points to something completely different and you have to rewrite everything.

Not to mention $4 is completely arbitrary and it's impossible to tell without prior knowledge which field it's referencing.

The original query `SELECT pwd, cmd WHERE exit-code != 0` is both descriptive and won't break with future updates.