常用漢字で、新字体の画数 < 異字体の画数のものを抽出 (元: https://gist.github.com/mandel59/8a7cd6cca228723e248d5fccbd41e69a )

jgquery.sh

#!/usr/bin/env bash

if ! command -v awk git grep sed wget &> /dev/null; then
  echo "install: awk, git, grep, sed, wget"
  exit 1
fi

[ -f joyo2010.txt ] || wget https://raw.githubusercontent.com/cjkvi/cjkvi-tables/master/joyo2010.txt
[ -f ucs-strokes.txt ] || wget https://raw.githubusercontent.com/cjkvi/cjkvi-ids/master/ucs-strokes.txt

if ! [ -d cjkvi-variants ]; then
  git clone --depth 1 https://github.com/cjkvi/cjkvi-variants
  rm -rf cjkvi-variants/.git
fi

: > res.csv
echo "常用,画数,異字体,画数" >> res.csv
sed 1d joyo2010.txt | awk '$0=$1' |
  while read i; do
    grep -r "/variant,${i}$" cjkvi-variants/*.txt | sed 's/^.*://' | sort | uniq
  done | awk -F, '{print$1" "$3}' |
  while read i j;do
    { # fast search
      grep "$j" ucs-strokes.txt | awk -F'[\t,]' '$0=$2" "$3'
      grep "$i" ucs-strokes.txt | awk -F'[\t,]' '$0=$2" "$3'
    } | xargs
  done |
  while read a b c d;do
    ((b<d)) && echo "${a},${b},${c},${d}"
  done >> res.csv

res.csv