Removing File Entries With sed

Shell Programming and Scripting
1

Hi

When running the following sed command from the command line I am able to
remove all entries with #username from the file and redirect to another file.

sed -e 's/#.*//' -e '/^$/d' namesonly > tt

mv tt > namesonly

When I use the same command in the following script I end up with the original
entries in the namesonly file. ie nothing has been removed.

 
echo "" > f-output.log
 
USERS=$(awk -F: '{print $1}' allusers)
for U in $USERS
 
do
 
finger ${U} >> f-output.log
 
awk '/^Logi/ {a=$3} /^Last log/ {print a, $6}' f-output.log > lastlogin
 
grep -v "2009" lastlogin > lastlogin_1 ; grep -v ":" lastlogin_1 | sort | uniq > newlog_1
 
cut -d' ' -f1 newlog_1 > namesonly
 
sed -e 's/#.*//' -e '/^$/d' namesonly > tt ; mv tt namesonly

done

Is there an error with this command or something else causing the problem?

Any help appreciated.

2

You are looping the script for each users. so make sure you append the strings in the "namesonly" file rather than overwriting it each time in the loop.

3

Thanks for your response....append the strings in the "namesonly" file
Can you provide an example... I tried the following which didnt work.
Do I need to change elsewhere in the script

sed -e 's/#.*//' -e '/^$/d' namesonly >> tt