Bash partial directory content

phual · 04-02-2010, 08:34 AM

I've got a fairly simple problem to be solved. I have a number of directories that contains files with different extensions. I want to run a script that gives different outputs for a given extension within a chosen directory.

E.g.

For directory 1, which contains:

about.xyz
file1.abc
file1.txt
file2.abc
file2.txt
file3.abc
file3.txt
support.xyz

output:

./rundifferentscript -files file1.abc file2.abc file3.abc -extra about.xyz support.xyz

Should be simple... can I figure it? Can I heck! Sorry for the dumb question - I just can't seem to wrap my head around bash.

Stuart

pixellany · 04-02-2010, 08:44 AM

I am not clear on what you need to do. Are you trying to sort the names? Or maybe you need to have the script take an action based on the name?

What you show as "output" just looks like how the command might be entered.

Do you have some kind of BASH tutorial?---If not, go to http://tldp.org and get the Bash Guide for Beginners

phual · 04-02-2010, 10:49 AM

I have a program (which I didn't write) which runs with a number of different input files. There are around six different file types that can be used, each file type being used differently by the script, and the script can utilise any number of each file type (usually one, but sometimes more). Each collection of files that go together are stored in their own directory (including one or more files of different file type), and when I run my script I want it to work on all files in that directory.

At the moment, I have to call the script by manually typing:

Code:

./rundifferentscript -files directory1/file1.abc directory1/file2.abc directory1/file3.abc -extra directory1/about.xyz directory1/support.xyz

As you can imagine, it gets tedious to do it this way. What I want to do is run a script where I specify the directory and it calls my other script with the parameters determined automatically, so that I only need to type:

Code:

./newscript directory1

I've tried a couple of times to work my way into bash, but I don't think that I'm going to manage it without a significant amount of time learning bash properly and motivation in the form of a more sustained need for the knowledge (my needs are simple and sporadic). My attempts to find answers myself never seem to work. For example, I've found that echo *.abc almost does the job (returning "file1.abc file2.abc file3.abc"), but I'd like to be able to do more manipulation than *.abc allows (such as prefix each file with a path).

Thanks for the link through. I'll need to sit down and give it some attention.

Stuart

IW2B · 04-02-2010, 10:49 AM

Hi Stuart,

For your script/command "rundifferent" to know what is -files and which is -extra you must either hard code the value that you want or accept it as a command parameter.

Sample Code:

Code:

#!/bin/bash
if [ $# -lt 1 ]; then
	 	echo "Usage $0 <extension>"
else
	echo "You specified an extension off $1"
	echo
	echo "files : Files matching $1 extension:"
	ls  -1 *.$1
	echo
	echo "extra : Files not matching $1 extension"
	ls -1  | grep -v \.$1
fi

Sample output:

Code:

$ sh ./rundifferent.sh 
Usage ./filelist.sh <extension>

$ sh ./rundifferent.sh abc
You specified an extension off abc

files : Files matching abc extension:
file1.abc
file2.abc
file3.abc

extra : Files not matching abc extension
about.xyz
file1.txt
file2.txt
file3.txt
rundifferent.sh
support.xyz

$ sh ./rundifferent.sh txt
You specified an extension off txt

files : Files matching txt extension:
file1.txt
file2.txt
file3.txt

extra : Files not matching txt extension
about.xyz
file1.abc
file2.abc
file3.abc
rundifferent.sh
support.xyz

$ ls -1
about.xyz
file1.abc
file1.txt
file2.abc
file2.txt
file3.abc
file3.txt
rundifferent.sh
support.xyz

This is just a quick script to show how I understood your question, I have not fully tested it.

Hopefully it will give you an idea/help you.

Regards
Ian

pixellany · 04-02-2010, 11:38 AM

Quote:

I've tried a couple of times to work my way into bash, but I don't think that I'm going to manage it without a significant amount of time learning bash properly and motivation in the form of a more sustained need for the knowledge (my needs are simple and sporadic).

I am going to divulge the secrets to learning BASH:
One command at a time
man -k <keyword>
man <command>
trial and error

catkin · 04-02-2010, 12:39 PM

Quote:

Originally Posted by pixellany

I am going to divulge the secrets to learning BASH:
One command at a time
man -k <keyword>
man <command>
trial and error

Very good advice; so much of using bash is automating execution of commands; to do that successfully you need to know a) what the commands are b) how to use them and c) what they output.

bash as a programming language is relatively small (but non-intuitive) so getting started with it is relatively easy (mastering its nuances may be more than many of us can do in a life time!).

If you have used C++ in a Windows environment, the situation is analogous; C++ itself is relatively simple while the Win32 interface libraries are immense and complex. Mastering the nuances of C++ may be easier than bash, though.

PTrenholme · 04-02-2010, 12:55 PM

And you might want to run pinfo find with special attention to the "Action" section. (Install the pinfo command if you haven't already done so. In my opinion it provides a much better interface to the info files.)

My point is that you may not need to use a bash script to accomplish your goal.

phual · 04-03-2010, 08:36 AM

Ian, interesting as it is the code you produced doesn't quite capture the problem fully due to the way in which it outputs the information (a new line per result).

From the general responses it does look as though I'm going to have to do it the hard way and do lots of reading.

Thanks anyway.

Stuart

PTrenholme · 04-04-2010, 02:31 PM

Using Ian's code, here's one way to do it:

Code:

#!/bin/bash
if [ $# -lt 1 ]; then
  echo "Usage $0 <extension>"
  exit 0
fi
# For testing, just copy the temp file to stdout.
# For "real," use something like tmp=/tmp/sdf.sh
tmp="/dev/stdout"
echo '#!/bin/bash' >$tmp
echo -n ". $PWD/rundifferentsript -files " >$tmp
for f in $(ls  -Q *.$1)
do
  echo -n "$f"' ' >$tmp
done
echo -n "-extra " >$tmp
for f in $(ls -Q --ignore='*.'"$1")
do
  echo -n "$f"' ' >$tmp
done
echo >$tmp
# To run the temp file, uncomment the following (the "rm" is for housekeeping).
#chmod u+x $tmp
#. $tmp
#rm -f $tmp

And here's a sample output from my bash scripts directory:

Code:

$ ./Ian.sh sh
#!/bin/bash
. /home/Peter/Scripts/bash/rundifferentsript -files "Convert to ext4.sh" "Ian.sh" "run-mozilla.sh" -extra "child" "Emerald_replace" "main" "mount_all_images_in" "new_script" "RepairDamagedCD" "Uncompress_Books

Note that, instead of the echo -n command used to output the text without a new-line at the end, you could, instead, use a "\" as the last character in file line and get something like this:

Code:

#!/bin/bash
. /home/Peter/Scripts/bash/rundifferentsript -files "Convert to ext4.sh" \
"Ian.sh" \
"run-mozilla.sh" \
-extra "child" \
"Emerald_replace" \
"main" \
"mount_all_images_in" \
"new_script" \
"RepairDamagedCD" \
"Uncompress_Books \

(You'd also need an "extra" echo after you echoed the last "extra" file name to terminate the command input.)