=====File management=====



==== Using the find command ====


<code>
find . -name '*httpd*'
</code>


Searches in current directory (.) and deeper for file or directory with httpd in name.


<code>
find /www -ctime -1
</code>

Searches directory ''**www**'' for files whose status was changed less than 24 hours ago.


<code>
find . -amin +1500
</code>

Searches current directory for files that were accessed more than 1500 minutes (or 25 hours) ago.


=== Excluding a directory from find ===

<code>
 find /home/ -path '/home/onno' -prune -o -name '*.txt' -ls
 11829831    4 -rw-------   1 ilke     ilke         1317 Oct 23 22:46 /home/ilke/.mozilla/firefox/vkuuxfit.default/cookies.txt
</code>

Searches in /home directory and beneath for .txt files, except in ''/home/onno'' directory.



==== The locate Command ====

If you already know the filename but not the file's location, use ''locate''. Example:

<code>
# locate libMagickCore.so.1
/usr/src/ImageMagick-6.4.8-3/magick/.libs/libMagickCore.so.1.0.0
/usr/src/ImageMagick-6.4.8-3/magick/.libs/libMagickCore.so.1
/usr/local/lib/libMagickCore.so.1.0.0
/usr/local/lib/libMagickCore.so.1
</code>

Apparently, ''locate'' relies on an index (database) which you can update as follows:

<code>
updatedb &
</code>

==== Other Tricks ====

<code c>
cat iptables-rules | less
</code>


Shows contents of file //iptables-rules// 





<code c>
chmod 600 iptables-rules 
</code>


Change access rights for file //iptables-rules//



Explanation:

4 - read

2 - write

1 - execute



user/owner - group - world



So, for the example given above, only the user/owner may read/write this file. 






====Copying the contents of a directory====


Do ''**man cp**'' and ''**man mv**'' to check out the copy and move commands. That said, it took me a while to figure out that the wildcard for files is NOT *.* but simply * under Linux. So, copying the contents of a directory goes like this:



<code c>
[root@1038 public_html]# cp -R ../dummy-3.7.0/* .
</code>


This means: copy the contents (*) of "dummy-.3.7.0 into the current (.) directory. The -R option (recursive) ensures that any subdirectories will be copied along.




====Creating symbolic links====
Unix / Linux supports symbolic links, which makes it a very flexible OS. Symbolic links enable you to use a subtree in another subtree without actually copying any directories. Issue this command:



<code c>
ln -s target_dir symlink_name
</code>


Add the full path if necessary.









====Creating short text files====
Linux uses filters such as ''**sort**'' to filter standard input. But you can also use a very simple filter, one which does not filter at all, to view the contents of a file:



<code c>
cat myfile.txt |less
</code>


What's more, you can redirect the standard input using cat to a file:



<code c>
[onno@1038 onno]$ cat > data.txt
This is a test!
</code>

Use Ctrl-D to save the file and get back to the shell prompt.

If you use the ''**>**'' redirect symbol, any data the file may have previously contained, is overwritten. Use  ''**>>**'' to append the data to the end of the file: ''**cat >> data.txt**''.



====Search and Replace in Text Files====


Use find and sed to search and replace strings in multiple text files.

<code>
find . -name '*.txt' -exec sed -i -e 's/find/replace/g' {} \;
</code>

You can use sed for multiple replacements, too

<code>
find . -name '*.txt' -exec sed -i -e 's/find1/replace1/g' -e 's/find2/replace2/g' -e 's/find3/replace3/g' ... {} \;
</code>

Use control+D to end the input.

Info taken from [[http://snippets.dzone.com/posts/show/4645|this forum]].

=== Find Pattern First ===
Take a look at what's going to be replaced first:

<code>
find . -exec grep "find1" '{}' \; -print
</code>

Or, simpler:

<code>
find . | xargs grep 'string' -sl
</code>
NOTE 20160414: didn't work for me anymore.

As explained [[http://www.liamdelahunty.com/tips/linux_find_string_files.php|here]]:

  * The -s is for summary and won't display warning messages such as grep: ./directory-name: Is a directory

  * The -l is for list, so we get just the filename and not all instances of the match displayed in the results.

**Or, even simpler:**
<code>
grep -Ril "text-to-find-here" .
</code>

  * R: Recursive
  * i: ignore case
  * l: "show the file name, not the result itself"
  * .: (the dot) start searching in the current directory



=== Rename Files Recursively and Replace Strings inside Files Recursively ===

This is a summary of the information provided earlier.

<code>
find . -iname '*valid*' -exec rename 's/valid/mentor/i' {} +
find . -name '*' -exec sed -i -e 's/valid/mentor/g' -e 's/Valid/Mentor/g' {} \;
</code>






====Applying patches to files using the patch utility====
Use the ''**patch**'' utility to apply patches, as contained within ''**.diff**'' files, to other files. Example:



<code>
[root@1038 public_html]# patch -p1 -i ../patches/moodle-1.6-patch.diff
patching file lib/javascript.php
patching file lang/en_utf8/moodle.php
Hunk #1 succeeded at 197 (offset 1 line).
patching file course/format/topics/format.php
patching file course/lib.php
</code>




====Making backups with tar====
Use the archiving and compressing utility tar to backup your files. Here's an example where we backup the contents of a webroot dir to a tar file:



<code c>
tar -zcpf tekstenweb20060824.tar public_html
</code>


Explanation:

  *z: zip
  *c: create
  *p: same permissions
  *f: make a file (instead of using the standard output)












====Renaming Files====

Simplest:
<code>
mv background_scifi.gif background.gif
</code>

=== Bulk Renaming ===
I have found the perl program ''**rename**'' to be an excellent tool for this. Its basics are explained on the '[[http://tips.webdesign10.com/how-to-bulk-rename-files-in-linux-in-the-terminal|Webmaster Tips]]' site.

Example:

<code>
rename -n 's/\_scifi//' *
</code>

This replaces the string "scifi" with the "" string (i.e. an empty string), for testing purposes. To do the real thing, simply omit the  ''-n'' parameter (or replace it with ''-v'' to see the actual results).

== Recursive Renaming ==
Here's another use for ''rename'', recursively:
<code>
find . -iname '*_en.html.erb' -exec rename 's/\_en.html.erb$/\.en.html.erb/i' {} +
</code>
This means: rename all ''*_en.html.erb'' to ''*.en.html.erb'' (replace the underscore by a dot), in the current directory and all subdirectories.

See also [[linux:scripts#recursive_version]] for more examples (replacing spaces in file names).


====Recursive Copying from an FTP Site====

How do you copy multiple embedded directories at once from an ftp site? Here's some advice: [[http://forums.devshed.com/ftp-help-113/recursive-mget-with-command-line-ftp-37472.html|forums.devshed.com]].

It comes down to: ''wget -r ftp://username:password@domain.com/the/path/to/directory''

If that does not work, try this:

<code>
ncftpget -R -v -u onno -p mysecretpassword the.ftp.host ~/Downloads/destination /the/source/directory
</code>

  * -R means recursive
  * -v means verbose (show everything)

====Recursive Copying TO an (S)FTP Site====

<code>
lftp -c "open -u USER_NAME,PASSWORD sftp://HOST:/TARGET_DIRECTORY ; mirror -R /SOURCE_DIRECTORY"
</code>





====Recursive Removal of Files ====

Combine ''find'' with ''rm'' to recursively delete files. To feed the results of ''find'' to ''rm'', pipe ''find'' to ''xargs''. This is explained on [[http://en.wikipedia.org/wiki/Xargs|wikipedia.org]].

Example:

<code>
find . -name \*~ | xargs rm
</code>
Delete all files in current directory and subdirectories which end in ''~''.

For files with spaces in their names:
<code>
find . -name "*.mobi" -print0 | xargs -0 rm
</code>

Another example:

<code>
find /home/onno/docs -name \*.bak | xargs rm
</code>
Delete all files in docs and subdirectories which have the file extension ''.bak''

Alternative:

<code>
find . -name "FILE-TO-FIND"-exec rm -rf {} \;
</code>


==== Print Files in a Batch ====

If you want to print all files in a directory, use the cli command ''lpr''.

First you have to find out what your printer is called exactly. This is especially important if you have more than one printer installed. Use the command ''lpstat'':

<code>
lpstat -p -d
printer hp-LaserJet-1300 is idle.  enabled since Thu 28 Jul 2011 11:41:06 AM CEST
	ready to print
system default destination: hp-LaserJet-1300
</code>

To find out what options are available for your printer, use the command ''lpoptions'' and specify your printer:

<code>
lpoptions -p hp-LaserJet-1300 -l
PageSize/Media Size: *Letter Legal Executive Statement A4 C5 C6 DL COM10 Monarch
ColorModel/Color Model: *Gray Black
StpColorPrecision/Color Precision: *Normal Best
(... etc ...)
</code>

Finally, to print all pdf documents in the current directory, for instance, issue the ''lpr'' command. WARNING: you will not be asked a confirmation, your printer will get to work immediately.

<code>
lpr -P hp-LaserJet-1300 -o media=A4 *.pdf
</code>


==== Copy Same File to Multiple Directories ====

To copy the same file to multiple subdirectories with the same structure, use the ''find'' command:

<code>
cd ~/php
touch delete_me_please.txt
find ./*/public_html/backup/moodle2 -maxdepth 1 -type d -exec cp ~/php/delete_me_please.txt {}/ \;
</code>

This will copy the file ''delete_me_please.txt'' from your php directory into all subdirectories matching the path ''./*/public_html/backup/moodle2''.

This is very nice if you want to copy over a patched file to multiple instances of the same system, say Moodle.

And to delete your test file:

<code>
find ./*/public_html/backup/moodle2/delete_me_please.txt -maxdepth 1 -type f -exec rm {} \;
</code>




==== Excluding directories from tar ====

<code>
tar --exclude 'local/soda/.git' --exclude 'local/soda/docs' -zcpf localplugin_soda.20121031.tar.gz local/soda
</code>