I have a folder in /opt/backup in which folders are created every day. In order to save space I would like to gunzip all folders that are older than 2 days.
I don't want to create one single zip file but rather zip each folder on its own, with the name preserved. I have tried:
#!/bin/bash
# Backup files
files=($(find /opt/backup/ -mtime +"2"))
for files in ${files[*]}
do
echo $files
tar cvfz backup.tar.gz $files
done
But all this does is creating a single zip file, I would like each folder separately.
The script will run every 2 days at 02:00 in the morning. How do I write this script, please?
You are making it too complicated. You should find directories that are old enough and simply tar zip those.
find /opt/backup/ -mtime +"2" -type d -exec tar cvfz backup.tar.gz {} \;
This will look for all directories (-type d) and execute a certain command on them (tar cvfz backup.tar.gz {}). In which {} is a placeholder for the directory found.
If you want to preserve the name of the dir, simply use {} a second time:
find /opt/backup/ -mtime +"2" -type d -exec tar cvfz {}.tar.gz {} \;
Note that no quotes are required around {} as special chars will be handled well inside find's exec.
Related
How can I bulk-zip folders in subdirectories without including the parent folder in the zip archives? I have a folder structure like this:
folder01
folder02
file01
file02
When I run:
find . -type d -name "folder02" -exec zip -r '{}'.zip '{}' \;
I get "folder02.zip" which always extracts its contents into a parent folder "folder01". How can I prevent this? For me it creates useless parent folder structures when extracting these archives anywhere else.
Using some simple bash:
find . -type d -name "folder02" -exec bash -c 'cd "$(dirname "{}")"; zip -r "$(basename "{}")".zip "$(basename "{}")"' \;
I have a folder with a number of subfolders. Each of the subfolders consists of several files without extension.
I need to add extension .cel to each file in subfolders.
How can I do it using bash?
find to the rescue:
find /your/folder -type f -exec mv {} {}.cel \;
Explanation: find obtains all files inside the /your/folder structure. From all the results obtained, it performs the mv command. It makes the file XXX to be moved to XXX.cel, which is another way of renaming it.
If you have rename then using that with find should do the trick:
find . -type f -exec rename -v 's/$/\.cel/' {} \;
So I have some folder
|-Folder1
||-SubFolder1
||-SubFolder2
|-Folder2
||-SubFolder3
||-SubFolder4
Each subfolder contains several jpg I want to zip to the root folder...
I'm a little bit stuck on "How to enter each folder"
Here is my code:
find ./ -type f -name '*.jpg' | while IFS= read i
do
foldName=${PWD##*/}
zip ../../foldName *
done
The better would be to store FolderName+SubFolderName and give it to the zip command as name...
Zipping JPEGs (for Compression) is Usually Wasted Effort
First of all, attempting to compress already-compressed formats like JPEG files is usually a waste of time, and can sometimes result in archives that are larger than the original files. However, it is sometimes useful to do so for the convenience of having a bunch of files in a single package.
Just something to keep in mind. YMMV.
Use Find's -execdir Flag
What you need is the find utility's -execdir flag. The GNU find man page says:
-execdir command {} +
Like -exec, but the specified command is run from the subdirec‐
tory containing the matched file, which is not normally the
directory in which you started find.
For example, given the following test corpus:
cd /tmp
mkdir -p foo/bar/baz
touch foo/bar/1.jpg
touch foo/bar/baz/2.jpg
you can zip the entire set of files with find while excluding the path information with a single invocation. For example:
find /tmp/foo -name \*jpg -execdir zip /tmp/my.zip {} +
Use Zip's --junk-paths Flag
The zip utility on many systems supports a --junk-paths flag. The man page for zip says:
--junk-paths
Store just the name of a saved file (junk the path), and do not
store directory names.
So, if your find utility doesn't support -execdir, but you do have a zip that supports junking paths, you could do this instead:
find /tmp/foo -name \*jpg -print0 | xargs -0 zip --junk-paths /tmp/my.zip
You can use dirname to get the directory name of a file/directory it is located in.
You can also simplify the find command to search only for directories by using -type d. Then you should use basename to get only the name of the subdirs:
find ./*/* -type d | while read line; do
zip --junk-paths "$(basename $line)" $line/*.jpg
done
Explanation
find ./*/* -type d
will print out all directories located in ./*/* which will result in all subdirs of directories located in the current dir
while read line reads each line from the stream and stores it in the variable "line". Thus $line will be the relative path to the subdir, e.g. "Folder1/Subdir2"
"$(basename $line)" returns the only the name of the subdir, e.g. "Subdir2"
Update: add --junk-paths to the zip command if you do not want the directy paths to be stored in the zip filde
So a little check, I finally got something working:
find ./*/* -type d | while read line; do
#printf '%s\n' "$line"
zip ./"$line" "$line"/*.jpg
done
But this create un archive containing:
Subfolder.zip
Folder
|-Subfolder
||-File1.jpg
||-File2.jpg
||-File3.jpg
Instead I fold like it to be:
Subfolder.zip
|-File1.jpg
|-File2.jpg
|-File3.jpg
So I tried using basename and dirname in differnet combination...Always got some error...
And just to learn how to: what if I would like the new archive to be created in the same root directory as "Folder"?
Ok finally got it!
find ./* -name \*.zip -type f -print0 | xargs -0 rm -rf
find ./*/* -type d | while read line; do
#printf '%s\n' "$line"
zip --junk-paths ./"$line" "$line"/*.jpg
done
find . -name \*.zip -type f -mindepth 2 -exec mv -- '{}' . \;
In first row I simply remove all .zip files,
Then I zip all and in the final row I move all zip to the root directory!
Thanks everbody for your help!
I want to copy just the files i've created/edited today into a separate directory "changeset" whilst maintaining their directory structure
I came up with the following script
cd ./myproject/
find ./* -mtime -1 -daystart -exec cp {} ../changeset/{} \;
The drawbacks of the above is that directories aren't created and the copy throws an error.
I've manually gone into ../changeset/ and create the folder structure until the command runs without errors.. but thats a little tedious.
Is there a simple solution to this?
find * -mtime -1 -daystart -print0 | cpio -pd0 ../changeset
cpio is an old, oddball archival program that is occasionally the best tool for the job. With -p it copies files named on stdin to another directory. With -d it creates directories as needed.
I've found another solution which isn't as elegant as John's but which isn't reliant on cpio, which i dont have.
cd ./myproject/
# Create all directories
find ./* -type d -exec mkdir ../changeset/{} \;
# Copy files
find ./* -mtime -1 -daystart -exec cp {} ../changeset/{} \;
# Delete empty directories, run this several times because after moving a child the parent directory needs to be removed
find ../changeset/ -type d -empty -exec rmdir {} \;
I've so far figured out how to use find to recursively unzip all the files:
find . -depth -name `*.zip` -exec /usr/bin/unzip -n {} \;
But, I can't figure out how to remove the zip files one at a time after the extraction. Adding rm *.zip in an -a -exec ends up deleting most of the zip files in each directory before they are extracted. Piping through a script containing the rm command (with -i enabled for testing) causes find to not find any *.zips (or at least that's what it complains). There is, of course, whitespace in many of the filenames but at this point syntaxing in a sed command to add _'s is a bit beyond me. Thank for your help!
have you tried:
find . -depth -name '*.zip' -exec /usr/bin/unzip -n {} \; -exec rm {} \;
or
find . -depth -name '*.zip' -exec /usr/bin/unzip -n {} \; -delete
or running a second find after the unzip one
find . -depth -name '*.zip' -exec rm {} \;
thx for the 2nd command with -delete! helped me a lot..
just 2 (maybe helpful) remarks from my side:
-had to use '.zip' instead of `.zip` on my debian system
-use -execdir instead of -exec > this will extract each zip file within its current folder, otherwise you end up with all extracted content in the dir you invoked the find cmd.
find . -depth -name '*.zip' -execdir /usr/bin/unzip -n {} \; -delete
THX & Regards,
Nord
As mentioned above, this should work.
find . -depth -name '*.zip' -execdir unzip -n {} \; -delete
However, note two things:
The -n option instructs unzip to not overwrite existing files. You may not know if the zip files differ from the similarly named target files. Even so, the -delete will remove the zip file.
If unzip can't unzip the file--say because of an error--it might still delete it. The command will certainly remove it if -exec rm {} \; is used in place of -delete.
A safer solution might be to move the files following the unzip to a separate directory that you can trash when you're sure you have extracted all the files successfully.
Unzip archives in subdir based on the file name (../file.zip -> ../file/..):
for F in $(find . -depth -name *.zip); do unzip "$F" -d "${F%.*}/" && rm "$F"; done
I have a directory filling up with zipped csv files. External processes are writing new zipped files to it often. I wish to bulk unzip and remove the originals as you do.
To do that I use:
unzip '*.zip'
find . | sed 's/$/\.zip/g' | xargs -n 1 rm
It works by searching and expanding all zip files presently in the directory. Later, after it finishes there are potentially new unzipped new files mixed in there too that are not to be deleted yet.
So I delete by finding successfully unzipped *.csv files, and using sed to regenerate the original filenames for deletion which is then fed to rm via the xargs command.