For diff on whole rows, use you could use Linux' CLI tools, in particular join(1) which is very similar to SQL's [LEFT/RIGHT] JOIN clause. Another option is the

For diff on whole rows, use you could use Linux' CLI tools, in particular join(1) which is very similar to SQL's [LEFT/RIGHT] JOIN clause. Another option is the combination of sort(1) and uniq(1); with proper choice of options, and, in some cases, repeating the file with old or new data, you can get: cat old.csv old.csv new.csv | sort | uniq -u # only the added/changed records cat old.csv new.csv new.csv | sort | uniq -u # only the ol d/deleted records cat old.csv new.csv | sort | uniq -d # only the unchanged records