Re: Count file lines in a directory tree

You mean this?

find -name '*.p[lm]' -print0 | xargs -r0 cat | wc -l
[download]

:-)

Update: I was using the unnecessarily longwinded find $ -name '*.pm' -o -name '*.pl' $

Update: to make this output the number of files, the easiest approach is a bit of Perl:

find -name '*.p[lm]' -print0 \
| perl -00000pe'++$a;END{print STDERR "Files: ",$a||0,", lines: "}' \
| xargs -r0 cat | wc -l
[download]

Makeshifts last the longest.

Comment on Re: Count file lines in a directory tree Select or Download Code

Replies are listed 'Best First'.
Re^2: Count file lines in a directory tree by sauoq (Abbot) on Nov 09, 2005 at 02:12 UTC
That's a useless use of... xargs. `find -name '*.p[lm]' -exec cat {} \; \| wc -l` [download] :-) -sauoq "My two cents aren't worth a dime.";	[reply] [d/l]
Re^3: Count file lines in a directory tree by Aristotle (Chancellor) on Nov 09, 2005 at 02:25 UTC
Congratulations, if you have 10,000 matching files, that line makes you a `cat` herder. `;-)` And a herder of useless `cat`s, to boot… If you’re going to do that, you can just do `find -name '.p[lm]' -exec wc -l {} \; \| awk '{ sum+=$1 } END { print +sum }'` [download] Moves a bit less data around. It also makes it trivial to count the number of files, as GrandFather’s code does: `find -name '.p[lm]' -exec wc -l {} \; \ \| awk '{ sum+=$1; ++num } END { print sum "lines in" num "files" }'` [download] Nice catch on the glob. `:-)` The `xargs` in mine is quite justified, so I don’t run one process per matched file. The `cat` is the easiest solution to the problem that if too many files for a single commandline are matched, `xargs wc` would run the `wc` multiple times, reporting multiple disjunct totals. This way, `xargs` runs multiple `cat`s (but very few in total), but `wc` always run exactly once. Makeshifts last the longest.	[reply] [d/l] [select]