Inserting a line with special characters (',`,$ etc) into a file using BASH - bash

Question

Hello I need help with some bash script.
I need to insert the below line into a file at a specific line number
su - test -c '$HOME/scripts/auto start' >> /home/test/auto_start.`date +%h_%d_%y`.log 2>&1
In the bigger problem statement, the line to be inserted at a specified line number & into the file, all are dynamic & need to be fetched from a csv file.
The csv file looks like:
$ cat test_file.csv
TARGET_NODE,TARGET_FILE,TARGET_LINE,NEW_LINE
10.10.10.10,test_file.csv,2,su - test -c '$HOME/scripts/auto start' >> /home/test/auto_start.`date +%h_%d_%y`.log 2>&1
Have tried using sed in many ways, but failed. We can't use python due to system limitations.
The simple way to test it:
seq 4|sed 's/<LINE_NUMBER>/<NEW_TEXT>/'
So we tried using:
seq 4|sed 's/2/'su - test -c '$HOME/scripts/auto start' >> /home/test/auto_start.`date +%h_%d_%y`.log 2>&1'/'

Answer 1

You didn't say if you wanted the new line printed before or after or instead of the existing line so here's all 3:
$ seq 4 | awk -F, 'NR==FNR{line[$3]=$4; next} FNR in line{print line[FNR]} 1' test_file.csv -
1
su - test -c '$HOME/scripts/auto start' >> /home/test/auto_start.`date +%h_%d_%y`.log 2>&1
2
3
4
.
$ seq 4 | awk -F, 'NR==FNR{line[$3]=$4; next} 1; FNR in line{print line[FNR]}' test_file.csv -
1
2
su - test -c '$HOME/scripts/auto start' >> /home/test/auto_start.`date +%h_%d_%y`.log 2>&1
3
4
.
$ seq 4 | awk -F, 'NR==FNR{line[$3]=$4; next} FNR in line{$0=line[FNR]} 1' test_file.csv -
1
su - test -c '$HOME/scripts/auto start' >> /home/test/auto_start.`date +%h_%d_%y`.log 2>&1
3
4
The above will work using any awk in any shell and for any string from that CSV on every UNIX box

Answer 2

If you have no system limitations for using perl :
#!/usr/bin/env bash
seq 4 | perl -pe '
BEGIN {
open CSV, shift; # Open test_file.csv
<CSV>; # Ignore first line
# Get line number and command
($n, $c) = <CSV> =~ /^(?:[^,]+,){2}([^,]+),(.*)/;
}
{ s/.*/$c/ if ($. == $n); } # Replace line with number $n
' test_file.csv

Answer 3

Just properly quote the line and use different s command separator because / is used in replacement string... escape special regex characters in replacement string, like & or \.
$ seq 4 | sed 's~2~su - test -c '\''$HOME/scripts/auto start'\'' >> /home/test/auto_start.`date +%h_%d_%y`.log 2>\&1~'
1
su - test -c '$HOME/scripts/auto start' >> /home/test/auto_start.`date +%h_%d_%y`.log 2>&1
3
4
with awk:
seq 4 | awk -vvar='su - test -c '\''$HOME/scripts/auto start'\'' >> /home/test/auto_start.`date +%h_%d_%y`.log 2>&1' 'NR==2{$0=var}1'
I discourage the usage of backticks `. Do not use them. Use $(...) instead.