So, I'm just starting with APL. I was thinking, it's possible to use APL in shell as part of a pipeline receiving a stream? We can write `awk 'program' file`, maybe there is any apl implementation that could work that way?
notice that APL isn't really that useful for processing non-numeric data & friends, the APL glyphs won't render in the shell prompt unless you use an APL font for your terminal. for a lot of normal unix-y/awk-y problems, the traditional toolbox of sed/awk/perl does the job well because of how specialised these tools are, while APL is somewhat vague, general and more multipurpose than, say, AWK.
AWK automatically loops, splits you input into the chunks, figures out in a "smart" way where to take input from, has conveniently available regexes, ...
Apart from rendering, typing the glyphs would be a problem, I can imagine prototyping something in a small dataset on the apl repl/editor session, and then copy the sentence for using in shell, or use a apl -f myscript file. In shell I would apply it to the complete dataset.
what matters at least to me is the convenience. i'd rather have quick awk/perl solution that does the job and move on to the rest of the problem than perfect this small and often insignificant bit. that's also the mindset of sysadmins i knew, etc...
You may try just forget about "shell" and think the APL session is your shell, and at least ⎕sh can call external programs via shell in the session. @jacksonbenete