Junk Drawer

For all those little papers scattered across your desk

The Fields Extracter You Always Wanted

11 Sep 2019 in Blog

Ever awk‘d a bunch of fields out of a pipeline? Tired of typing single quotes, dollar signs, and prints, when you just want the fields? Me too.

See github for the most up-to-date version.

Update 15th September 2019: Most of this can be done with cut(1) via the -f flag. I might rewrite the script to take advantage of this, but preserve the same functionality.

Example

# ls -l | fields 1 9
total
-rw-r--r-- 404.md
-rw-r--r-- Gemfile
-rw-r--r-- Gemfile.lock
-rw-r--r-- LICENSE
-rw-r--r-- README.md
-rw-r--r-- Session.vim
-rw-r--r-- _config.yml
drwxr-xr-x _data/
drwxr-xr-x _drafts/
drwxr-xr-x _includes/
drwxr-xr-x _layouts/
drwxr-xr-x _posts/
drwxr-xr-x _sass/
drwxr-xr-x _site/
drwxr-xr-x _writings/
drwxr-xr-x assets/
-rw-r--r-- index.md
drwxr-xr-x pages/
-rwxr-xr-x serve*
drwxr-xr-x sitemap-styles/
lrwxr-xr-x sitemap.xsl@
drwxr-xr-x subs/
-rwxr-xr-x tags-list*

How?

The script (below) is a wrapper around awk(1): we just generated the code we would normally type by hand, and feed it to awk(1).

If you’re concerned about timing to generate code, replacing eval with echo and running fields {1..100} takes 0.012s on my machine. (If you’re pulling out 100 fields, I have questions for you…)

I’m considering replacing eval with exec, but we will have to see.

The script

#! /usr/bin/env bash

set -euo pipefail

log() {
  printf '%s\n' "$@"
} >&2

usage() {
  cat <<DOG
usage: $0 field...
DOG
}

die() {
  local ex="${1:-1}"
  exit "$ex"
}

usage_and_die() { usage && die; }

generate_awk_code() {
  local fields=("$@")
  case $# in
    0) return ;;
  esac
  printf "'{ print "
  # from the beginning to the second to last
  for field in "${fields[@]:0:${#fields[@]}-1}"; do
    printf '$%d, ' "$field"
  done
  # last
  printf '$%d' "${fields[${#fields[@]}-1]}"
  printf " }'"
}

main() {
  eval awk "$(generate_awk_code "$@")"
}

main "$@"

Tags:

Categories: Blog

Load Comments
Previous Next
Back to posts