Filter lines based on length using sed and grep

How to locate lines longer than 100 and smaller than 150 characters using sed and grep

2 Answers

Using sed command

sed -nr '/^.{101,149}$/p' infile

Breaking down:

• -n: no printing

The -n option will prevent to print anything unless an explicit request to print is found (^.{101,149}$ "Lines longer than 100(>=101) and smaller than 150 characters(<=149)"). The /p flag used as one way to turn printing back on.

• -r: -r (--regexp-extended)

The -r enables the Extended regular expressions. replace -r with -E if your sed doesn't recognize it or you will need to escape the curly braces if none of them supported in your version of sed. Then the command would be sed -n '/^.\{101,149\}$/p' infile.

• The ^ is just the anchor for beginning of line and the $ for end of line.

• A single . matches any single character

So, you got it, with sed -nr '/^.{101,149}$/p' infile, we are looking for a line from start to end of it which has or contains our defined rule.

_{This is a way to duplicate the function of grep with sed e.g: sed -n '/.../p' :)}

Using grep command

grep -E '^.{101,149}$' infile

• -E this is the same as sed's -r option but for grep command. So if you don't want to use is just escape the curly braces.

Using awk command

awk 'length($0)>100 && length($0)<150' infile

• length return the length of the line. In awk, the $0 specifies the whole line. So if the length of line was between 100 and 150, will be print out.

Another way of doing this using bash:

while read l; do nc=$(<<< "$l" wc -c); [ $nc -ge 101 ] && [ $nc -le 149 ] && echo "$l"; done < file

Expanded into a script:

#!/bin/bash
while read l; do nc=$(<<< "$l" wc -c) if [ $nc -ge 101 ]; then if [ $nc -le 149 ]; then echo "$l" fi fi
done < file