#!/bin/bash
Version=2.01
 Myname="${0##*/}"

:<<'DOC'
= reorder - reorder columns in a file

= Synopsis
reorder [options] --fields=list [file]	

== Options
-h,--help	print short help and exit
-H,--Help	print full documentation and exit
-V,--version	print version and exit
-f,--fields=X	set list of list of comma-separated field(range)s to X
-s,--squeeze	squeeze off trailing empty fields
-t,--tab=X	set column separator to X; default: the tab character
-u,--undefined=X	set contents for undefined fields to X; default: empty
-i,--inplace	change file inplace


= Description
reorder reorders the columns in a file (standard input by default) according to
the list of comma-separated field numbers and/or field ranges given, where a
field range is two numbers separated with a hyphen. Reverse ranges are allowed.

By default, columns are separated by tabs, unless a different separator is
specified with the |--tab| option. Columns are counted starting at 1. Ranges of
columns may be specified with hyphens.

Unavailable columns are replaced with empty ones or, if the |--undefined| option
is used, with that options' argument. So if you specify the fifth column in a
2-column file, you end up with a 5-column file.

Output is written to standard output, unless the the |--inplace| option is used,
which causes the output to be written to the original file.

= Examples
Switch the second and third columns in a tab-separated file:

        reorder --fields=1,3,2 file

Reverse the three columns of a space-separated file:

        reorder -t' ' -f3-1 file

Move columns 7 through 9 after the last column in an 11-column
tab-separated file:

        reorder -f1-6,10,11,7-9 file

Put two empty columns between the columns of a 2-column file:

        reorder -f1,3,4,2 file

Insert an empty column before the first in a file with a variable number
of columns, with a maximum of 12, say:

        reorder -f20,1-12 file

This will result in a 13-column file with trailing empty fields!

The same, but replace undefined fields with '-' and remove trailing empty
fields, and replace the original file with the output:

        reorder -siu- -f20,1-12 file

= Author
[Wybo Dekker](wybodekker@me.com)

= Copyright
Released under the [GNU General Public License](www.gnu.org/copyleft/gpl.html)
DOC

REd='\e[38;5;9m'
    die() { local i; for i; do echo -e "$Myname: $REd$i"; done 1>&2; exit 1; }
helpsrt() { sed -n '/^= Synopsis/,/^= /p' "$0"|sed '1d;$d'; exit; }
helpall() { sed -n "/^:<<'DOC'$/,/^DOC/p" "$0"|sed '1d;$d'|
            less -P"$Myname-${Version/./·} (press h for help, q to quit)";exit; }

:<<'DOC' #----------------------------------------------------------------------
= handle_options
synopsis:	 handle_options "$@"
description:	handle the options.
globals used:	 Myname Version
globals  set:	 args
DOC
#-------------------------------------------------------------------------------
handle_options() {
   local options
   options=$(getopt \
      -n "$Myname" \
      -o hHVIst:u:if: \
      -l help,Help,version,squeeze,tab:,undefined:,inplace,fields: \
      -- "$@"
   ) || exit 1
   eval set -- "$options"
   sep=$'\t' inplace=false undef='' squeeze=false
   while [ $# -gt 0 ]; do
      case $1 in
      (-h|--help)       # print short help and exit
			helpsrt
			;;
      (-H|--Help)       # print full documentation and exit
			helpall
			;;
      (-V|--version)    # print version and exit
			echo $Version
			exit
			;;
      (-f|--fields)     # set list of list of comma-separated field(range)s to X
			fields="$2"
			shift 2
			;;
      (-s|--squeeze)    # squeeze off trailing empty fields
			squeeze=true
			shift
			;;
      (-t|--tab)        # set column separator to X; default: the tab character
			sep="$2"
			[[ ${#sep} == 1 ]] ||
			die "--tab option: column separator must be one single character"
			shift 2
			;;
      (-u|--undefined)  # set contents for undefined fields to X; default: empty
			undef="$2"
			shift 2
			;;
      (-i|--inplace)    # change file inplace
			inplace=true
			shift
			;;
      (-I)              instscript "$0" ||
			   die 'the -I option is for developers only'
			exit
			;;
      (--)              shift
			break
			;;
      (*)               break
			;;
      esac
   done
   args=( "$@" )
}

handle_options "$@"
set -- "${args[@]}"

[[ -z $fields  ]] && die 'you must use the --fields option'
case $# in
(0) $inplace && die "you may not use the --inplace option on standard input"
    file=$(mktemp -t "$Myname.XXXXXXXXXX")
    trap 'rm -f "$file"' 0 1 2 15
    cat >"$file"
    ;;
(1) file="$1"
    ;;
(*) die "too many arguments"
esac 
cols=()
list=$(sed -e '
  s/,/ /g
  s/\([0-9]\+\)-\([0-9]\+\)/{\1..\2}/g
' <<<"$fields")
list=$(eval "echo $list")
[[ $list =~ - ]] && die "open ranges like n- or -n are not allowed"

for i in $list; do
   [[ $i =~ ^[1-9][0-9]*$ ]] || die "illegal column number «$i»"
   (( j=i-1 ))
   cols+=( "$j" )
done
[[ ${#cols[@]} == 0 ]] && die "I need at least one column number"

tmp=$(mktemp -t "$Myname.XXXXXXXXXX")
trap 'rm -f "$tmp"' 0 1 2 15

while IFS=$'\2' read -ra x; do 
   z=''
   (( n=${#x[@]}-1 ))
   for i in "${cols[@]}"; do 
      if (( i>n )) ; then p="$undef"; else p="${x[$i]}"; fi
      z+="$p$sep"
   done
   printf '%s\n' "${z%$sep}" >> "$tmp"
done < <(sed -e "s/$sep$/$sep$sep/" < "$file" | tr "$sep" '\2')
#        ^^^ if a row ends with a separator, we want that to mean that it ends
#        with an empty field, not with an undefined field. Since the read
#        function discards a final IFS, we double it here.

if $squeeze; then 
   sed -i -e "s/$sep\\+$//" "$tmp"
fi

if $inplace; then
   mv -f "$tmp" "$file" 
else 
   cat "$tmp"
fi