How to make a list of txt ogg or other files?

[Deonis]

Hello
Do somebody know how to make a list of all mp3 ogg or any other files in linux and save it as csv. Or, do somebody know a program or python script which allow you to do that by just pointing at the location with this files??

Any help?

[William Haller]

find / -name '*.ogg' -type f >mylist.csv

comes to mind...

[Deonis]

Thanks for reply !!!
It's look like a very handy method ! Is there a way to put file size and name of the file separately ??? Or I want too much!?

[William Haller]

You can always get more creative with the find command

Add a --exec ls -la {} \;

to get the file information, followed by a cut command based on columns to isolate the file size and name.

Do a man find and man cut for more information.

[jpollard]

I would suggest using perl... but for a special reason.

Just because a file has a given extension doesn't mean the file really is that type. I would tend to
use the perl find module to list all files, and then verify the file type with the "file" utility.

This also allows you to get the directory entry via stat.

The file module includes a file type function (using magic numbers to identify file types).

The major advantage of using perl is that only one process is needed for the entire thing.
You also have a possibility to to this in a multi-threaded manner - where speed is necessary -
I've been able to scan a filesystem with over 40 million files in about 45 minutes that way.

[Deonis]

finally this work just as I expected:
sudo find /media/ -name '*.ogg' -type f -ls >mylist.csv
Thanks to everybody who were trying to help !!!

[Deonis]

# this is the python script which is doing about a same thing except no formatting. And
# unfortunately I am not that good at python so I can not really figure out how
# to make similar output of "find" command. if somebody will figure out please
# post here

import os, time
ff = raw_input("type a name of the folder to scan:  ")
# ff = '/user/folder/'
allfiles = []
subfiles = []

for root, dirs, files in os.walk(ff):
     for f in files:
         if f.endswith('.mp3') or f.endswith('.ogg'):
             allfiles.append(os.path.join(root, f))
             if root != ff: 
                  subfiles.append(os.path.join(root, f))   
#stats = os.stat(root)
#lastmod_date = time.localtime(stats[8])
#date_file_tuple = lastmod_date, files
#files.append(date_file_tuple)
#files.sort()
#files.reverse() 
#print "%-60s %s" % ("filename:", "last modified:")
#for file in files:
                   #  folder, file_name = os.path.split(file[1])
                    # file_date = time.strftime("%m/%d/%y %H:%M:%S", file[0])
                    # print "%-60s %s" % (file_name, file_date)   
print "*"*80          
print "allfiles=", allfiles
print "-"*112

[jpollard]

The only difference I see is that find prints the filename as discovered instead of putting it in an array. Or
forks/execs a process to do something with it - such as "ls" -- find . -depth -name "..." -o -name "..." -exec ls -l '{}' ';'

The main disadvantage of that is that a new process is created for each file (lots of fork/exec). What I was suggesting
was to use the language capabilities to do a file content check the way the "file" utility does. It reads the first 16 bytes
(or so) to check against the "magic" (man magic) list to determine content. Takes longer than an extension list,
but if files are not always ending in the default extension, it works quite well.