Using sed, how can I delete all instances of a string within a range of lines?

Question

I know that

sed '/match/ d' file

deletes all lines with matches, and

sed '1,3 s/match//g' file

deletes all matches in the first 3 lines.

But how do I delete all lines with matches in first 3 lines?

If possible, give a solution with only one sed call (no piping).

Answer 1

You could combine the two:

sed '1,3{/match/d;}' file

This would delete lines containing match in the specified address range, i.e. in lines 1-3 in the example above.

Answer 2

If you like to test awk, it would be:

awk '/match/ && NR<=3 {next} 1' file

Answer 3

awk -v string="foo" 'NR<=3 && index($0,string) {next} 1' file