Using a for loop I am able to get the following log file below. I would like o change the format a bit as is indicated lower down. I have looked at awk and sed but somehow is not getting any joy. I am a bit new to ksh scripting. Please assist.

This is the format of my current log file after the script has run.

/edi/iTracTEST/sessions/otherreceive/nohup.out @ 08:19:46 from 63 - 63
/edi/iTracTEST/sessions/payroll/nohup.out @ 08:19:46 from 52395 - 63
/edi/iTracTEST/sessions/payrollmsg/nohup.out @ 08:19:46 from 56931 - 63

As from the first record/line.
The path and file name is already in a variable from a the for loop in the script.
The "@" is echoed.
The time is coming from the date command.
The "from" is echoed.
The first 63 is the initial file size in bytes as from a variable in the for loop.
The "-" is echoed.
The second 63 is the current/new file size as from a variable in the for loop.

I would like the log file to look like this. !!!Notice I have swapped the sizes!!!

/edi/iTracTEST/sessions/otherreceive/nohup.out 08:19:46 63 63
/edi/iTracTEST/sessions/payroll/nohup.out 08:19:46 63 52395
/edi/iTracTEST/sessions/payrollmsg/nohup.out 08:19:46 63 56931

This just make the log file more readable.
Here is some of the code, thus far.

sess_path=/edi/iTracTEST/sessions/
ls -pl /edi/iTracTEST/sessions/ | awk '$1 ~ /^d/ $9 ~ /\/$/' | awk '{print $9}' > `pwd`/temp/`date +"%C%y%m%d"`.tmp
for nohup_loc in $(cat `pwd`/temp/`date +"%C%y%m%d"`.tmp)
do
nohup_loc2=$sess_path$nohup_loc"nohup.out"
if test -s $nohup_loc2
then
cur_fsize=`ls -l $nohup_loc2 | awk '{print $5}'`
echo !!! Reset nohup.out file at `date` !!! > $nohup_loc2
new_fsize=`ls -l $nohup_loc2 | awk '{print $5}'`
echo $nohup_loc2 @ `date +%H:%M:%S` from $cur_fsize - $new_fsize >> `pwd`/log/`date +"%C%y%m%d"`_nohup_reset.log
fi
done