Bash remove exact matching variables and search for string ignoring case

Question

First Query:

i have a text file username.txt contains

user1
user2
user3

My Bash file is as follows > test.sh

#!/bin/bash
# REMOVE user which donot have any COMBLIMIT
num=0
cat /tmp/username.txt | while read users
do
num=$[$num+1]
USERID=`echo $users | awk '{print $1}'`
SRVID=`mysql -uxxx -pxxx -e "use radius; SELECT srvid FROM radius.rm_users WHERE rm_users.username = '$USERID';" |awk 'FNR == 2 {print $1}'`
COMBLIMITCHECK=`mysql -uxxx -pxxx -e "use radius; SELECT limitcomb FROM rm_services WHERE srvid = '$SRVID';" |awk 'FNR == 2 {print $1}'`
if [[ $COMBLIMITCHECK -eq "1" ]]; then
echo "$USERID have Quota limit = 1"
else
echo "$USERID have no Quota Limitations = 0 , So removing it from the users list."
sed -i '/\$USERID\>/d' /tmp/username.txt
fi
done

sed -i '/\$USERID\>/d' /tmp/username.txt >> This command in bash is not removing the exact matched username. I want to delete line which contains specific username (variable).

Second Query:

File contains following lines

cOke is bad 
diet-COKE 
cokeisgood

How can i cat this file and search for any word which contains "coke" in it regardless of lower or upper case.

Answer 1

sed -i '/\$USERID>/d' /tmp/username.txt

You have enclosed your sed expression in ' single quotes. This will stop the expansion of $USERID and cause it to be used as a string literal. You should enclose the sed expression in " double quotes to allow the variable expansion.

File contains following lines

The -i switch to grep ignores case.

You can also use I (capital i) in a sed expression for (this may be a GNU sed thing)

echo "cOke is bad" | sed -n "/coke/Ip"