Simple Script Problem - reordering

[Russ Hancock_1]

Does anyone know how I can reorder a file, so the first line of the input file becomes the last line of the output file, and so on through out the input file...

I don't want it sorted, just reversed...

Thanks in advance.
Russ.

[U.SivaKumar_2 · Accepted solution]

Hi,
If you know total number of lines of that file
you can do this for file of 900 lines.
#tail -n 900 myoriginalfile > invertedfile

#man tail for more details.

regards,
U.SivaKumar

[Steve Steel]

Hi

nl -ba -s"~" file|sort -r|cut -f2-256 -d"~"


Use umlaut or something else for field definition see man


steve Steel

[Ralph Grothe]

Hi Russ,

this is yet another task for Perl ;-)

But beware, this is a resource greedy solution as it requires the size of your file from memory stack, since we store the whole file in an array first.
So you shouldn't do this with really big files:

perl -e '@tmp = reverse <>;print @tmp' /path/to/your/file


Unfortunately, if you have big files, something more elaborate is needed, which I imagine would require seek().

HTH
Ralph

[Russ Hancock_1]

Here's what I did in the end.
Thanks All.

INPUT=$1
OUTPUT=$2

rm -f $OUTPUT

LINES=`wc -l $INPUT|awk '{print $1}'`

i=0

while test i -lt $LINES
do
i=`expr $i + 1`
echo " $i \c"
tail -n $i $INPUT | head -n 1 | tee -a $OUTPUT
done

[john korterman]

Hi,
you could try this rather simple solution:

#!/usr/bin/sh

# reverse order of inputlines in $1

PREVIOUS_FILE=/tmp/prev
if [ -w $PREVIOUS_FILE ]
then
rm $PREVIOUS_FILE
fi
touch $PREVIOUS_FILE

CURRENT_FILE=/tmp/curr
if [ -w $CURRENT_FILE ]
then
rm $CURRENT_FILE
fi
touch $CURRENT_FILE

cat $1| while read CURRENT
do
echo $CURRENT > $CURRENT_FILE
cat $PREVIOUS_FILE >> $CURRENT_FILE
cat $CURRENT_FILE > $PREVIOUS_FILE
done
cat $CURRENT_FILE



regards,

John K.

[V. V. Ravi Kumar_1]

hi,

try this
cat -n |sort -r|cut -f2 >

regds

[H.Merijn Brand (procura]

When using perl, you can use http://search.cpan.org/search?dist=File-ReadBackwards which is not that resource greedy

[Peter Godron]

Russ,
you could also try:
grep -n ^ input-file | sort -kr1n,1n | awk -F':' '{print $2}'

[Aashish Raj]

#! /usr/bin/ksh
i=1
cat file.txt|while read xx
do
buf[$i]=`echo $xx`
i=$((i+1))
done
v=$((i-1))
while true
do
if [ $v -ge 1 ];then
echo ${buf[$v]}
v=$((v-1))
else
exit
fi
done

Thanks
AR

[James R. Ferguson]

Hi Russ:

Here's one way:

# awk '{array [NR]=$0}; END{for (i=NR;i>0;i--) print array [i]}' inputfile > outputfile

Regards!

...JRF...

[Reinhard Burger]

HI
I would prefer the solution f U.Sivakumar
because it follows my most important approach : Keep it simple

just use:
tail -10000000000 oldfile > newfile

it's a silly way just to use a very very big number for tail but IT WORKS and you get what you want.
As long as your file does not have that big amount of lines you will get the file "oldfile" listed into "newfile" in reverse order.

[Scott Williams_5]

Hi Russ,

Try this:

ed yourfile <g/^/m0
w
EOD

[Ralph Grothe]

Hey procura,

thanks for the hint towards File::ReadBackwards module from CPAN

http://search.cpan.org/doc/URI/File-ReadBackwards-0.97/ReadBackwards.pm

Haven't stumbled yet over this one.
Of course this is *much* better than what I suggested as a quick one-liner.

I will definitely install this one.
It's always worth to dig the CPAN first

:-)

[Shannon Petry]

For you two who said to use the tail command to do this, I have a suggestion for you. Know what you are saying before giving answers on the forums!

Tail DOES NOT reverse order display of files, simply takes the last "n" lines of the file to stdout.

I.E. (yes I clipped the last two lines of cat)
cat /etc/hosts
127.0.0.1 localhost
10.37.161.12 sh0su012 tr0su010 loghost

tail -2 /etc/hosts
127.0.0.1 localhost
10.37.161.12 sh0su012 tr0su010 loghost


Doh, they are identical....
Look at the other suggestions for your answers and ignore the suggestions of using the tail command.

Regards,
Shannon

[Rodney Hills]

Also "tail" does have a limitation on how many lines it can process.

If you say tail -1000000, tail does a sequential read and has to keep the last 1000000 lines in memory. Sometime tail doesn't allocate enough memory, and I've seen where it just drops off the remaining lines.

My 2 cents...

-- Rod Hills

[Shannon Petry]

Sadly the guy who said tail for this got 9 points and a magic answer....guess the user didnt test it, just had faith.. ;(

[Reinhard Burger]

Ok Guys
You got me !!!
Next time i'm hopefully a little bit mor carefull
:;-)

[Suhas_2]

Hi Guys,
Here are my 2 cens.
On Solaris, IBM-AIX and DIGITAL a mere
"tail -r /path/to/file/filename"
would do the job. On HP-UX "-r" option is not available.

Rgds..
Suhas.

[U.SivaKumar_2]

Hi Shanon,
I do agree that. I posted a theory that i have not tried practically. I regret , If i have hurted you all.

regards,
U.SivaKumar

[U.SivaKumar_2]

And shannon,

"For you two who said to use the tail command to do this, I have a suggestion for you. Know what you are saying before giving answers on the forums!"

To err is human, You can be more polite in pointing out other's mistake.

regards,
U.SivaKumar

[Russ Hancock_1]

Thanks Guys, Enough already!

Shannon, maybe 9 points was excessive for the tail command, but it did point me to the solution of using tail to pick the last line and then looping backup one line at a time.. See my reply earlier.

So tail didn't work as suggested... but it did prove an answer.... and that what mattered to me.