cut, tr, and paste in Linux: Small Tools Add Up

Some text jobs are too small for awk and too fiddly for an editor: take the second column, lowercase everything, join two files side by side. That is the exact territory of cut, tr, and paste — three small tools that each do one obvious thing.

cut: columns out #

Take the first and third field

$ cut -d: -f1,7 /etc/passwd
$ cut -c1-10 long-lines.txt

-d sets the delimiter, -f picks fields, and -c slices by characters when there is no delimiter at all. The passwd example prints users and their shells — a classic quick audit.

tr: characters across #

tr translates or deletes single characters, always reading stdin:

Lowercase, strip digits, squeeze repeats

$ echo "Version 2.0.1" | tr 'A-Z' 'a-z'
$ echo "call 555-1234" | tr -d '0-9'
$ echo "too    many    spaces" | tr -s ' '

That last one, squeeze, is the tidy-up pass for messy command output. Uppercase a whole file in a pipe:

Uppercase a headline

$ head -1 notes.txt | tr 'a-z' 'A-Z'

paste: streams together #

paste joins lines side by side — the inverse of cut, and the fastest way to build a two-column table from two lists:

Join names and scores with a tab, then a comma

$ paste names.txt scores.txt
$ paste -d, names.txt scores.txt > pairs.csv

Tool	One job
`cut`	Extract fields or character ranges
`tr`	Translate, squeeze, or delete characters
`paste`	Merge lines across files
`rev`	Reverse each line (honourable mention)

A small assembly line #

Chained, the three replace a spreadsheet import:

Build clean CSV from raw dumps

$ paste -d, <(cut -f1 data.tsv) <(cut -f3 data.tsv) | tr -s ' ' > clean.csv

Process substitution — the <(...) form — feeds each cut as a file; it is shell trickery worth an evening of practice on its own.

Counting what you cut? sort and uniq finish that job. The full terminal notes series covers each hand-off like this one.