How do I get the last non-empty line of a file using tail in Bash?

Question

How do I get the last non-empty line using tail under Bash shell?

For example, my_file.txt looks like this:

hello
hola
bonjour
(empty line)
(empty line)

Obviously, if I do tail -n 1 my_file.txt I will get an empty line. In my case I want to get bonjour. How do I do that?

@ghostdog74 great solution https://stackoverflow.com/a/2641745/658497 — Noam Manos, Oct 17 '21 at 09:02

score 39 · Answer 1 · answered Apr 14 '10 at 16:13

39

Use tac, so you dont have to read the whole file:

tac FILE |egrep -m 1 .

answered Apr 14 '10 at 16:13

Jürgen Hötzel

18,997
3
42
58

@vault not on Sierra, 2016. – gsamaras Nov 28 '16 at 17:05
1

@gsamaras try swiping from right to left on the touchbar – vault Nov 29 '16 at 12:16
Then the error that `gtac` is not found won't be shown @vault . . . ;) – gsamaras Nov 29 '16 at 15:32
@gsamaras the touchbar is also lick-able . And, by the way, this worked for me on CEntOS 7.5 (vagrant box `bento/centos-7.5`). – Nathan Basanese Nov 20 '18 at 19:51
3

Why use `egrep` instead of `grep`? There is nothing "extended" in the regular expression. – jarno Mar 15 '19 at 10:31
Besides `egrep` is deprecated according to the manual page of `grep` nowadays. – jarno Dec 12 '20 at 16:23

score 29 · Accepted Answer · answered Apr 14 '10 at 23:20

29

You can use Awk:

awk '/./{line=$0} END{print line}' my_file.txt

This solution has the advantage of using just one tool.

answered Apr 14 '10 at 23:20

Hai Vu

37,849
11
66
93

9

Downside: Needs to read the WHOLE file, and assign EACH LINE to awk variables -> This can be quite CPU and IO intensive. – Jürgen Hötzel Apr 16 '10 at 19:40
2

Yes, the solution is simple, but far from being efficient. An efficient solution would open the file, seek to the end and scan backward. – Hai Vu Apr 16 '10 at 23:24

score 26 · Answer 3 · edited Jun 07 '16 at 12:56

26

How about using grep to filter out the blank lines first?

$ cat rjh
1
2
3


$ grep "." rjh | tail -1
3

edited Jun 07 '16 at 12:56

Andreas Louv

46,145
13
104
123

answered Apr 14 '10 at 15:55

RichieHindle

272,464
47
358
399

7

or, instead of 'cat rjh | grep "." | tail -1', use 'grep . rjh | tail -1' – amertune Apr 14 '10 at 17:03
could someone explain the grep "." part ? – Ab5 Oct 30 '15 at 06:46
1

In `grep "."`, that period is regex, meaning "any character". In regex a blank line has a beginning (^), an end ($ or \n), but nothing in the middle, so grepping for the period only returns lines that contain characters. E.g., not blank. – Aaron R. Dec 12 '17 at 18:25
1

Simple > efficient. Bonus: `... | grep \. | tail -1` one char shorter! – hyperpallium Nov 26 '18 at 03:46

score 6 · Answer 4 · answered Dec 26 '13 at 16:07

6

Instead of tac you can use tail -r if available.

tail -r | grep -m 1 '.'

answered Dec 26 '13 at 16:07

kalu

61
1
1

1

Which versions of `tail(1)` have the `-r` option? The one in GNU coreutils doesn't have it, cf. http://debbugs.gnu.org/18808 – Josip Rodin Jul 04 '16 at 10:20
The issue with this solution is that if the file you're scanning is a shell script file then it will only filter the echo output rather than the actual file contents, try 1) creating a script file with a few simple echo commands; 2) pasting the full file path into the console and entering " | tail -1" at the end; notice that it will only output your last echo command's output rather than the entire echo command... – 255.tar.xz Jun 03 '19 at 21:58

score 4 · Answer 5 · answered Apr 14 '10 at 23:26

4

if you want to omit any whitespaces, ie, spaces/tabs at the end of the line, not just empty lines

awk 'NF{p=$0}END{print p}' file

answered Apr 14 '10 at 23:26

ghostdog74

327,991
56
259
343

score 2 · Answer 6 · answered Apr 26 '16 at 15:18

2

If tail -r isn't available and you don't have egrep, the following works nicely:

tac $FILE | grep -m 1 '.'

As you can see, it's a combination of two of the previous answers.

answered Apr 26 '16 at 15:18

Sophie

304
10
12

`egrep` is the same as `grep` just instead of needing to put `... | grep "my[regex]Expr"` you only need to put `... | egrep my[regexExpr]` it's a little more complex that that as you may run into a syntax issue, however a few escape characters should do the trick. – 255.tar.xz Jun 03 '19 at 22:01

pormulsys · Answer 7 · 2022-05-20T11:10:28.057

I had problems using other solutions, so I made this.

First, get last 25 lines, assuming at least 1 is not empty. Filter out empty lines, and print out the last line.

 tail -n25 file.txt | grep -v "^.$" | tail -n 1

One major advantage this has, you can show more than 1 line, just by changing the last 1 to, lets say, 5. Also, it only reads last 25 lines of the file.

If you have huge amounts of empty lines, you might want to change the 25 to something bigger, repeating until it works.

jarno · Answer 8 · 2020-04-02T12:45:58.250

0

Print the last non-empty line that does not contain only tabs and spaces like this:

tac my_file.txt | grep -m 1 '[^[:blank:]]'

Note that Grep supports POSIX character class [:blank:] even if it is not documented in its manual page until 2020-01-01.

File may contain other non-visible characters, so maybe using [:space:] may be better in some cases. All space is not covered even by that, see here.

edited Apr 02 '20 at 12:45

answered Mar 15 '19 at 10:34

jarno

787
10
21

1

If you want to add more lines, change the number after `-m` and add `|tac` to the end to get it back in correct order – Eric Wolf Dec 11 '20 at 23:04

score 0 · Answer 9 · answered May 20 '21 at 14:47

0

Using ag way:

cat file.txt | ag "\S" | tail -1

answered May 20 '21 at 14:47

ー PupSoZeyDe ー

1,082
3
14
33

How do I get the last non-empty line of a file using tail in Bash?

9 Answers9

Linked

Related