Using sed to Print Specific Lines in a File
This could be handy when a stack trace in a large file points to a specific line, now you can see those lines quickly.
Every once in a while I’m doing contract work where I end up having a massive 200,000+ line file, maybe it’s a SQL dump or something else but if a command that processes that file doesn’t work and a stack trace says there’s a problem on line 134,810 sometimes I don’t want to open the whole file in an editor. I might even be in a container where no editor is installed.
You can use sed -n 50p myfile to output just line 50 specifically, or if you
want to see a range of lines you can do sed -n 50,100p myfile to see those
lines. We use -n so sed doesn’t print the whole file and the p at the end
tells sed to print matching lines.
The range is inclusive so 5,8 would include lines 5, 6, 7 and 8.
You can do more advanced ranges too, such as using regular expressions using
sed -n '/PATTERN_1/,/PATTERN_2/p' myfile. You can also do something from
let’s say line 10 to the end of the file with sed -n '10,$p' myfile where $
is the last line. Make note to use single quotes here otherwise your shell
will try to interpret $p as a variable and it will be empty.
You can even print multiple sets of lines and ranges too using ; to split
them up. For example sed -n '3,11p;35p' myfile would print lines 3-11 and
also line 35.
Of course if you have a small file you can easily cat it or even use less
for a decently small file but every once in a while using sed helps a lot!
It’s a versatile tool.
# Demo Video
Timestamps
- 0:17 – Real world use case
- 0:38 – Getting 1 line or a range of lines
- 1:07 – Regular expressions
- 2:01 – Different groups of lines
When was the last time you used sed to preview lines in a file?