@gumnos

Whoa, that's some #awk alphabet soup.

I'm gonna need to learn awk next, ain't I? XD

I actually started out doing it all in grep (pipe the words in the file to grep -vixfF /usr/share/dict/words, but it was actually twice as fast to do a positive grep per input word than one big negative grep.

I'm actually already removing duplicates via an awk command that I copypastad from the internet, but I'll play with your example, thanks! :D

@rl_dane yowzer, that grep-per-source-word is likely where your slowneess resides. You can abuse #awk a bit like

awk -F'[^[:alpha:]][^[:alpha:]]*' 'NR==FNR{d[$1]; next}{for (i=1; i<=NF; i++) if ($i && ! ($i in d)) {print $i; ++t}} END{exit (t>255 ? 255 : t) + 0}' /usr/share/dict/words your_file_to_spellcheck.txt

to read the dictionary once into a hash-map, then iterate over the lines+words in your file to see if any words are not present in the dictionary with O(1) lookup-time-per-word rather than O(N) lookup ☺

@linuxgal

For a pure #awk one-liner:

$ awk '{for (i=1;i<=NF;i++) if ((w=tolower($i)) ~ /^[a-z]{4}$/) ++a[w]}END{for (w in a) print a[w], w}' kjv.txt

That provides the frequencies too, but adjust the END block as you see fit 🙂