\1

On a popular Linux magazine, one of the writers asked people to come up with different ways to print the lines of a file in reverse order, ie like what the program tac does.

So here are some ways to do that, almost all of which are silly, useless, crazy, dangerous and inefficient. You have been warned.

Some can work with both standard input and a real file, some only work with a real file (the examples all use a file). The shell examples use some bash-specific features for convenience, although they could all be rewritten to use only standard constructs (except the one that uses arrays).

No explanations to avoid spoiling the fun (some are obvious though).

tac file

cat -n file | sort -k1,1rn | cut -f 2-

awk '{a[NR]=$0} END {for(i=NR;i>0;i--)print a[i]}' file

perl -e 'print reverse <>' file

perl -ne 'push @a,$_; print reverse @a if eof' file

sed '1!G;h;$!d' file

awk '{out = $0 s out; s = RS} END {print out}' file

i=0
while IFS= read -r arr[i]; do ((i++)); done < file
for((j=i-1;j>=0;j--)); do printf "%s\n" "${arr[$j]}"; done

sed ':a;$!{N;ba;}; s/$/\x1/; 
     :b; s/\n\{0,1\}\([^\n]*\)\x1\(.*\)/\x1\2\n\1/; /^\x1/!bb; 
     s/\x1\n//' file

i=0
while true; do
  if gawk -v n="$i" 'NR==n{exit}END{print;exit NR-1}' file; then
    break
  else
    i=$?
  fi
done

# NEVER do this
touch outfile
while IFS= read -r line; do
  { rm outfile; { printf "%s\n" "$line"; cat; } > outfile; } < outfile
done < file
cat outfile
rm outfile

i=0
while IFS= read -r line; do
  ((i++))
  printf "%s\n" "$line" > "$i".txt
done < file

while [ $i -gt 0 ]; do
  cat $i.txt
  ((i--))
done

rm *.txt

# same idea as above, unnecessarily extra-complicated...
shift_names() {
  n=$1
  while [ $n -gt 0 ]; do
    mv "$n.txt" "$((n+1)).txt"
    ((n--))
  done
}

i=0
while IFS= read -r line; do
  shift_names $i
  printf "%s\n" "$line" > 1.txt
  ((i++))
done < file

for((j=1;j<=i;j++)); do
  cat $j.txt
done

rm *.txt

{
printf '<?xml version="1.0" encoding="UTF-8"?>
<soapenv:Envelope xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/">
  <soapenv:Body>
    <a:sort
         soapenv:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" 
         xmlns:a="http://linuxtoolbox.kwfgrid.net">  
      <a:params>-k1,1nr</a:params>
      <a:value>'

recode utf8..xml < file | sed = | sed 'N;s/\n/ /' | awk -v ORS='&#xa;' 1

printf '</a:value>
    </a:sort>
  </soapenv:Body>
</soapenv:Envelope>'
} | curl -s -o - -d @- \
-H 'Content-Type: application/soap+xml;charset=utf-8' \
-H 'SOAPAction: ""' \
'http://www.gridworkflow.org:8080/linuxtoolbox/services/Sort' |\
xmlstarlet sel -T -t -v '//sortReturn' | perl -pe 's/^\d+ //'

If you can think of other ways (for example based on different ideas, more convoluted, or using other programming languages, etc.) contributions are welcome.

Update 02/01/2011 - here's another one:

LC_ALL=C awk -v x="0" '{y += length($0) + 1; x = y RS y FS x}END{print substr(x, index(x, RS) + 1)}' file | \
while IFS=" " read -r a b; do
  dd skip=$b count=$((a - b)) bs=1 2>/dev/null < file
done