How to Search for Files Recursively into Subdirectories

Question

I am trying to look for all XML files in a particular directory and all sub-directories (recursively) inside it.

ls -R *.xml is only listing files in the current directory. I am quite sure, the sub-folders themselves have several .xml files, but none are showing up.

Is this a configuration issue?

Solution 1

Try using Find

sudo find . -print | grep -i '.*[.]xml'

Solution 2

You can do it with find only:

find . -name '*.xml'

. is the current directory. If you need to search in another directory, replace . with the directory path.

Solution 3

Try this command:

ls -R | grep '.*[.]xml'

ls doesn't have options to filter the output. For that you would need to use pipe. This passes the output from ls to grep, which then filters them to show just the .xml files.

Solution 4

bash

Using globstar shell option, we can make use of recursive globbing ./**/*

bash-4.3$ shopt -s globstar
bash-4.3$ for i in  ./**/*.xml; do printf "%s\n" "$i" ; done
./adwaita-timed.xml
./bin/hw5/stuff/book/chapter42servletexample/build/web/META-INF/context.xml
./bin/hw5/stuff/book/chapter42servletexample/build/web/WEB-INF/beans.xml
./bin/hw5/stuff/book/chapter42servletexample/build/web/WEB-INF/web.xml

Perl

Perl has a module Find, which allows for recursive directory tree traversal. Within the special find() function, we can define a wanted subroutine and the directory that we want to traverse, in this example that's .. The one-liner in such case would be:

bash-4.3$ perl -le 'use File::Find; find(sub{-f && $_ =~ /.xml$/ && print $File::Find::name},".")' 
./adwaita-timed.xml
./CLEAR_DESKTOP/blahblah/hw5/stuff/book/jsf2demo/build/web/WEB-INF/beans.xml
./CLEAR_DESKTOP/blahblah/hw5/stuff/book/jsf2demo/build/web/WEB-INF/web.xml
./CLEAR_DESKTOP/blahblah/hw5/stuff/book/liangweb/build.xml

Python

While Perl has a whole module dedicated to recursive tree traversal, Python has a neat function walk() that is part of os module, and repeatedly returns tuple of topmost path, list of all subdirectories, and list of filenames. We can do the following:

bash-4.3$ python -c 'import os,sys; [ sys.stdout.write(os.path.join(r,i)+"\n") for r,s,f in os.walk(".") for i in f if i.endswith(".xml") ]' 
./adwaita-timed.xml
./CLEAR_DESKTOP/blahblah/hw5/stuff/book/jsf2demo/build/web/WEB-INF/beans.xml
./CLEAR_DESKTOP/blahblah/hw5/stuff/book/jsf2demo/build/web/WEB-INF/web.xml
./CLEAR_DESKTOP/blahblah/hw5/stuff/book/liangweb/build.xml

This might be far neater as a script:

#!/usr/bin/env python
import os,sys 
for r,s,f in os.walk("."): 
    for i in f: 
        if i.endswith(".xml") 
             print(os.path.join(r,i))

find

Other answers have mentioned find for recursive traversal, and that's the go-to tool for the job. What does need mention is the fact that find has multiple command line switches, such as -printf to print output in desired format, -type f to find only regular files, -inum to search by inode number, -mtime to search by modification date, -exec <command> {} \; to execute a particular command to process the file with passing file as argument ( where {} is standard find placeholder for current file) , and many others so please read the manpage for find.

View more solutions

Comments

[KeyC0de]

You can do ls -R | grep .xml

[Shamim Hafiz - MSFT]

is the sudo must, or it's there to ensure super user privileges?

[Mitch]

I let you decide. Sudo, No Sudo.

[don.joey]

Just out of interest. What is the advantage of find over ls -R?

[Mitch]

@don.joey This might help stackoverflow.com/questions/13830036/…

[Mostafiz Rahman]

Does it search for the required file recursively in the directory rooted at current directory. In my case it just checked in the current directory only, didn't check the subdirectory.

[KaeruCT]

@mostafiz, you need to quote the '*.xml' part. I'll edit my answer.

[Rodislav Moldovan]

@ShamimHafiz, sudo is useful if you need to search in a location where your current user does not have read right, I think this was the reason why sudo was included in command, just to be sure that you'll get all possible results

[Mostafiz Rahman]

Actually I searched for .php files in current directory. But it returned only .php files in current directory, didn't searched recursively in sub-directories. That's why I'm asking whether find command searches recursively or not.

[KaeruCT]

@mostafiz, the find command searches recursively. If you don't quote the parameter, I think your shell might do an expansion on the *, so it will match the files in the current directory.

[Mostafiz Rahman]

All right! May be I'd made some mistake. Now it is working perfectly!

[muru]

-1 for mixing find and grep, when find can do filtering using both regexes and globs, and not using find's -print0 and grep's -z when you do need to mix.

[user1767754]

What is the print statement for, when removing it or not, the result is the same.

[AdamO]

Anyway to get this to show the directory it came from?

[Christian]

@Mitch the image hosting site you are using is blocked from where I am. Can you please share what the images were showing...? :)

[Mitch]

@Christian Try Sudo, No Sudo.

[Kip]

you can use -regex or -iregex instead of -name if you want to use a regex.

[Iluvathar]

Mandatory link: Why not parse ls?

[wogsland]

Awesomesauce. Simple and effective!

[George Birbilis]

thanks, needed that in Raspbian, combined with |shuf -n 1 at the end to pick one file randomly from the recursive list (to play a random .mp3 file from a folder hierarchy if some motion sensor was on)