find length of longest line in all text files in a directory

Question

I know how to get length of the longest line in a text file with awk

awk ' { if ( length > L ) { L=length} }END{ print L}' file.txt

but how can I get the length of the longest line of all files in a directory?

lgeorget · Accepted Answer · 2014-08-18T15:33:51.947

13

The most straightforward solution is to concatenate all the files and pipe the result to your script:

cat ./* | awk '{ if ( length > L ) { L=length} }END{ print L}'

You can also pass directly several files to awk:

awk '{ if ( length > L ) { L=length} }END{ print L}' ./*

Of course, there can be some warnings if files are in fact directories but it should be harmless. You may have bigger problems with binary files because they don't have a concept of line. So, in order to be more specific, you can do something like

 awk '{ if ( length > L ) { L=length} }END{ print L}' ./*.txt

to match only the .txt files in the current directory.

And, as @G-Man stated in his comment, * won't match hidden files (starting with a dot). If you want those, use * .*.

edited Aug 18 '14 at 15:33

answered Aug 18 '14 at 14:58

lgeorget

13,914

Or, arguably, cat * .* | .... Or eliminate the useless use of cat and say awk '...' * .*. – G-Man Says 'Reinstate Monica' Aug 18 '14 at 15:07
@G-Man Indeed. I was editing my answer about that when you commented. :-) I said most straightforward because concatenating the files with cat is what seemed simpler at first. – lgeorget Aug 18 '14 at 15:09
1

use this modification to also output the content of the longest line: awk '{ if ( length > L ) { L=length ;s=$0 } }END{ print L,"\""s"\"" }' **/*.txt – python_kaa Aug 10 '21 at 17:02

Stéphane Chazelas · Answer 2 · 2015-09-11T08:55:14.013

If you want the max length per file, with GNU awk:

find . -type f -exec awk -v l=0 '
   length>l {l=length} ENDFILE{print FILENAME ":", l; l=0}' {} +

Or the one max length in all the files:

find . -type f -size +1c -exec cat {} + |
  awk -v l=0 'length>l {l=length}; END{print l}'

That assumes the files end in newline characters. If one file doesn't end in a newline character, then its last non-delimited line will be merged with the first line of the next file and possibly void your result.

-size +1c is an optimisation as text files that are empty or contain only one character have respectively 0 line and 1 empty line, so won't have the longest line.

They're text files, so they do end in newline characters (unless empty). — Gilles 'SO- stop being evil', Aug 18 '14 at 21:14

score 6 · Answer 3 · answered Aug 29 '14 at 09:12

6

With GNU wc:

cat *.txt|wc -L

-L prints the length of the longest line.

answered Aug 29 '14 at 09:12

Lri

5,223

1

GNU's -L option is horrible, it counts single tabs as being 8 bytes each – Harold Fischer Mar 25 '20 at 01:49

score 0 · Answer 4 · edited Feb 13 '15 at 15:06

0

Also with GNU wc (coreutils 8.4), it can handle multiple files

wc -L *.txt

lists the length of each individual file
as well as provide the longest length of all files

edited Feb 13 '15 at 15:06

Anthon

79,293

answered Feb 13 '15 at 14:43

Francis W

1

This is essentially what this other answer mentioning wc -L does, but your solution has the disadvantage that you have to wade through the max-length of all the other files first. Is there any advantage in that? – Anthon Feb 13 '15 at 15:05

find length of longest line in all text files in a directory

4 Answers4