Uncompress several tar.gz files inside several folders

Hi,

I need to uncompress sevral tar.gz files that are located insides sevral folders using a shell script.

Folder tree looks like this :

/folder/001
/folder/002
/folder/003

Inside each folder dossier (001,002,003) we can find 1 or several tar.gz files.

I've scripted something like this :

!/bin/bash
liste_fichier="/folder/*"
for fichier in $liste_fichier do
tar -xzvf $fichier/*.tar.gz --directory $fichier/
done
exit

Issue:
This scripts only works with one tar.gz file inside the folder.
When it finds more then one tar.gz file it crashes the message tar : <path/file.tar.gz> : not found in archive

Some other tests:
-I've tested the code below inside one folder(001) containing several tar.gz.file :

for i in *.tar.gz 
do 
tar -xzvf $i 
done

It works.

I try to run it inside my script without success. To many loops I guess

Any "good" ideas are welcome.
Thank you

Wold you be better with the find command? You can ask it to locate files of the name format you need and call a command for each one found.

Perhaps this will get you started:-

find /base/path/to/files -name "*.tar.gz" -print -exec tar -tzvf {} \;

Beware that if the files found contain full path archives, you might overwrite any file on your server. Additionally, if you are extracting, you might find that your current directory gets all the extracted files into it if they use relative paths. I have set the command to just list the contents for now so you can check what it will try to do.

I hope that this helps,
Robin

Hi,

That's the point, I must extract the file in the same folders.
Besides it doesn't work.
eRR mEssage : missing parameters for "-exec"

What is your operating system and shell used ?
Gnu find printf option can be of help if you operating system supports it.

Following should fill you requirement regardless of the paths used in archives (absolute or relative).

find /folder/00[1-3] -name '*.tar.gz' -printf 'tar -xvzf %h/%f -C %h\n' | sh

You can match folders with regex or input them along each other.

find /folder001 /folder002 /folder003 -name '*.tar.gz' -printf 'tar -xvzf %h/%f -C %h\n' | sh

More portable solution could include a awk program which will generate those commands instead of gnu find printf or some xargs magic possibly.
Be sure to get back, if you do not have gnu find.

Hope that helps.
Regards
Peasant.

@PEASANT

Hi,

This solution would be great for a small number of folders.
But unfortunnatelly "for me" I've something like 1000 sub-folders with tar.gz files to uncompress.

I cannot type the names of 1000 folders and subfolders in a script.

That's why my idea in the beggining was to accept a variable "$1"to store the path to the folders containing the tar.gz files
This way the script could be executed automatically for each folder,
(e.g.) below

!/bin/bash 
liste_fichier="/folder/$1/*" 
for fichier in $liste_fichier do 
tar -xzvf $fichier/*.tar.gz --directory $fichier/ 
done 
exit

Assuming that your code actually starts #!/bin/bash then you can actually simplify this to:-

#!/bin/bash

[ -z "$1" ] && echo "On doit donner un dossier." && exit

for ficher in /folder/$1/*
do
   echo ; echo "$ficher" ; sleep 1                                 # Remove this if you like, it is just to illustrate the loop
   tar -xzvf $fichier/*.tar.gz --directory $fichier/
done

The shell will expand the directory name with wildcard to make a long list for your for loop.

I've assumed French language for output from your use of ficher. Apologies if that is incorrect.

Does that help?
Robin

Hi Robin

That code is very good.

The issue is it doesn't work when there are more then one tar.gz file in the folder.

--> Lets suposse $1 = folder
--> The content of my folder "folder" is

[shellx] $ ls -l folder
total 16
drwxr-xr-x 2 shellx mrv 4096 11 juil. 14:36 001
drwxr-xr-x 2 shellx mrv 4096 11 juil. 14:37 002
drwxr-xr-x 2 shellx mrv 4096 11 juil. 14:37 003
drwxr-xr-x 2 shellx mrv 4096 11 juil. 14:34 004

--> sub-folders (001, 002, 003, 004) content

[shellx] $ ls  folder/*
folder/001:
CARDV022A0.tar.gz  CARDV02976.tar.gz  CARDV02A0.tar.gz  CARDV976.tar.gz

folder/002:
CARDV02976.tar.gz  CARDV976.tar.gz

folder/003:
CARDV022A0.tar.gz  CARDV02976.tar.gz  CARDV976.tar.gz

folder/004:
CARDV02976.tar.gz  CARDV976.tar.gz

-->I've set 2 loops in the new script

#!/bin/bash
liste_rep="folder/*"
folder="$(ls folder)"
echo "LISTE REP :" $liste_rep
        for rep in "$liste_rep"
        do
                echo "AFFICHE REP :"$rep
                #for fichier in "$(ls /folder/*)" # test
                for fichier in $liste_rep/*.tar.gz
                do
                        #cd $liste_rep
                        echo "FICHIER :" $fichier
                        tar -zxvf $fichier   --directory $liste_rep/
                done
        done
exit

-->Variables content verified

[presta2] $ echo $liste_rep
folder/001 folder/002 folder/003 folder/004
-(jeu. juil. 12 10:13:41)--(mrvpgsa001:~)-
[presta2] $ echo $folder
001 002 003 004

Resultat d'execution

LISTE REP : folder/001 folder/002 folder/003 folder/004  --->> folder liste correct
AFFICHE REP :folder/*  --->> because $rep = "$list_rep"??? and should be $list_rep without " " ?
FICHIER : folder/001/CARDV022A0.tar.gz  --> Var fichier is correctly set
tar: folder/002 : non trouv� dans l'archive ---> How is it ? if  --> tar -zxvf $fichier
tar: folder/003 : non trouv� dans l'archive
tar: folder/004 : non trouv� dans l'archive
tar: Arr�t avec code d'�chec � cause des erreurs pr�c�dentes
FICHIER : folder/001/CARDV02976.tar.gz
tar: folder/002 : non trouv� dans l'archive
tar: folder/003 : non trouv� dans l'archive
tar: folder/004 : non trouv� dans l'archive
tar: Arr�t avec code d'�chec � cause des erreurs pr�c�dentes
FICHIER : folder/001/CARDV02A0.tar.gz
tar: folder/002 : non trouv� dans l'archive
tar: folder/003 : non trouv� dans l'archive
tar: folder/004 : non trouv� dans l'archive
tar: Arr�t avec code d'�chec � cause des erreurs pr�c�dentes
FICHIER : folder/001/CARDV976.tar.gz

--> If uncomment line "#cd $liste_rep" inside second loop
--> we gett error message below

[presta2] $ ./decomp2.sh
LISTE REP : folder/001 folder/002 folder/003 folder/004
AFFICHE REP :folder/*
FICHIER : folder/001/CARDV022A0.tar.gz
tar (child): folder/001/CARDV022A0.tar.gz : la fonction open a �chou�: Aucun fichier ou dossier de ce type
tar (child): Error is not recoverable: exiting now
tar: Child returned status 2
tar: Error is not recoverable: exiting now
FICHIER : folder/001/CARDV02976.tar.gz
tar (child): folder/001/CARDV02976.tar.gz : la fonction open a �chou�: Aucun fichier ou dossier de ce type
tar (child): Error is not recoverable: exiting now
tar: Child returned status 2
tar: Error is not recoverable: exiting now

Any ideas are welcome

Please use code tags as required by rules.
Other then that, you still haven't provided your shell and operating system version.

Nobody expects you to write a 1000 names in a script.
That is why we have computers :slight_smile:

But you can probably match them with a regular expression.

We cannot help you form one, if you do not provide representative example of the directory list.

Output in french language also obfuscates things a bit.

Regards
Peasant.

Without fully understanding what going on here, the first thing to do would be drop the double quotes around $liste_rep in the first for loop, and replace $liste_rep with $rep in the second. Try that and report back.

1 Like

Sorry

---------- Post updated at 06:40 PM ---------- Previous update was at 06:35 PM ----------

Finally I found out the solution, Thank you Rudic

#!/bin/bash liste_rep=folder/* 
    echo "LISTE REP :" $liste_rep 
    for rep in $liste_rep 
##$liste_rep without "" double quotes
    do 
              echo "AFFICHE REP :"$rep 
              for fichier in $rep/*.tar.gz 
### use $rep instead of $liste_rep > varibale $liste_rep just list all folders  
### $rep shows folder/001 ... 003 
              do 
                   echo "FICHIER :" $fichier 
### $fichier shows complete path to tar.gz files in each sub-folder 
                   tar -zxvf $fichier --directory $rep/ 
### -- directory uncompress inside -> $rep = folder/001 with '/' so folder/001/...003/ 
             done 
    done 
exit
2 Likes