How to remove a given line from a file

I'm writing a BASH script to integrate our company's custom intranet with our postfix email server. In one situation, the script mu开发者_运维知识库st delete a line from the postfix virtual file which takes on the following form (for the most part): emailAddress username.

    xyx@abc.com xyz
    tuv@abc.com tuv
    lmn@abc.com lmn

my bash script needs to read through a file that contains user names (one on every line) and delete the corresponding line from the virtual file. So lets say it reads a line containing the username xyz; thus in the example below the variable $usr is storing the value 'xyz'.

touch virtual.tmp
cat virtual | while read LINE ; do
if [ "$LINE" != "$usr@abc.com $usr" ] ; then
echo "$LINE" >> virtual.tmp
fi
done
rm -rf virtual
mv virtual.tmp virtual

However this code doesn't work and it is probably not efficient, as all I want to do is delete a line based off of a username. Presumably I might not have to read the whole file.

---

Try sed with -i, like:

sed -i -e "/^$usr@abc.com/d" virtual

that should remove the line from virtual, no need to write to a temp file.

---

You have a file of usernames to delete, and a file (or stream) of records containing the username. Perfect situation for awk:

awk '
  FNR == NR {user[$1]++; next}
  !($2 in user) {print}
' users email_users > new_email_users