Trying to run a script that will fetch all directories, and files containing these directories, and logs data onto a .CSV file.
So, if I were to have structure like:
mainDir.dir -> [sub1.dir -> file01.png, sub2.dir -> file02.png]
, I would get a CSV of
dir; file
sub1; file01.png
sub2; file02.png
This is the script I currently have
for dir in */ .*/ ;
do
for entry in $dir
do
path="$entry"
empty=""
file="${$dir/$empty}"
echo -e "$dir;$file;" >> file.csv
done
done
find is useful for processing many files recursively.
Command
find . -type f -execdir sh -c "pwd | tr -d '\n' >> ~/my_file.csv; echo -n ';' >> ~/my_file.csv; echo {} | sed -e 's/^\.\///' >> ~/my_file.csv" \;
Note: make sure you do not give a relative path to the output CSV file. execdir changes the working directory (and that is what makes pwd work).
Breakdown
find . -type f find all files recursively starting here
-execdir sh -c "pwd | tr -d '\n' >> ~/my_file.csv; echo -n ';' >> ~/my_file.csv; For each file, execute in its directory pwd. Strip the newline and add directory name to output. Also add a semicolon, again with no newline.
echo {} | sed -e 's/^\.\///' >> ~/my_file.csv" \; Append filename to output. This time, leave newline, but by default find will place the ./ in front of the filename. The sed here removes it.
If you don't need more than one level deep, this seems to work
for i in **/*; do echo $i | tr / \; ; done >> file.csv
Related
I'm trying to loop through every file in a directory (including files in its subdirectories) and perform some action if the file meets an if-condition.
Part of my code is as follows:
for f in $direc/*
do
if grep -q 'search_term' $f; then
#action on this file
fi
done
However, this fails in the case of subdirectories. I would be very grateful if someone could help me out.
Thank you!
The -R option to grep will read all files in the directory tree including subdirectories. Combined with the -l option to print only the matching file names, you can use that to perform an action on each file that matches.
egrep -Rl pattern directory | while read path; do echo $path && mv $path /tmp; done
For example, that would print the file name and move the file to a different directory.
Find | xargs is the usual pattern I use, and has the advantage of not getting hung up on special characters in file names (spaces etc.) if you use the -print0 option of find.
find . -type f -print0 | xargs -0 -I{} sh -c "if grep -q 'search string' '{}'; then cmd-to-run '{}'; fi"
Yes because with this syntax, grep expect to process file(s) not directories. Minimal change to your script would be to test if $f is a file or not:
...
if [ -f "$f" ] && grep -q 'search_term' $f; then
...
In reality you would probably want to get list of files with patter match and act on those:
while read f; do
: #action on file file $f
done < <(grep -rl 'search_term' $direc/)
I've opted for getting the get the list of files through <(list) because piping it into while would cause the inside of your loop to run in another process (which could be a problem in particular if you expect any variable (changes) to be accessible from outside. And unlike simple for with `` it's not as as sensitive to what filenames you encounter (namely I have spaces in mind, this would still get confused by newlines though). Speaking of which:
while read -d "" f; do
: #action on file file $f
done < <(grep -rZl 'search_term' $direc/)
Nothing should be able to confuse that, as entries are nul character delimited and that one just must not appear in a file name.
Assuming no newlines in your file names:
find "$direc" -type f -exec grep -q 'search_term' {} \; -print |
while IFS= read -r f; do
#action on this file
done
I have been trying to recursively rename files AND folders with iconv without success, the files are correctly renamed but folders dont.
What I use for files is (works perfect):
find . -name * -depth \ -exec bash -c 'mv "$1" "${1%/*}/$(iconv -f UTF8 -t ASCII//TRANSLIT <<< ${1##*/})"' -- {} \;
What I tried for files AND folders (fail: Only rename folders):
find . -exec bash -c 'mv "$1" "$(iconv -f UTF8 -t ASCII//TRANSLIT <<< $1)"' -- {} \;
ORIGINAL problem:
I just want to bulk rename lots of files to make them "web friendly", thinks like removing spaces, weird characters and so on, currently I have
find . -name '*' -depth \
| while read f ;
do
mv -i "$f" "$(dirname "$f")/$(basename "$f"|tr -s ' ' _|tr -d "'"|tr -d ","|tr - _|tr "&" "y"|tr "#" "a")" ;
done
Is there any way to do the tr stuff above and the iconv at a single run? because I am talking around 300,000 files to rename, I would like to avoid a second search if possible.
If needed, I am working with Bash 4.2.24
Thanks in advance.
I think the following does everything you want in one pass.
# Update: if this doesn't work, use read -d '' instead
find . -print0 | while IFS= read -d '$\000' f ;
do
orig_f="$f"
# Below is pure bash. You can replace with tr if you like
# f="$( echo $f | tr -d ,\' | tr "$'&'#- " "ya__" )"
f="${f// /_}" # Replace spaces with _
f="${f//\'}" # Remove single quote
f="${f//-/_}" # Replace - with _
f="${f//,}" # Remove commas
f="${f//&/y}" # Replace ampersand with y
f="${f//#/a}" # Replace at sign with a
f=$( iconv -f UTF8 -t ASCII//TRANSLIT <<< "$f" )
new_dir="$(dirname $f)"
new_f="$(basename $f)"
mkdir -p "$new_dir"
mv -i "$orig_f" "$new_dir/$new_f"
done
The find command (no real options needed, other than -print0 to handle filenames with spaces) will send null-separated file names to the while loop (and someone will correct my errors there, no doubt). A long list of assignments utilizing parameter expansion removes/replaces various characters; I include what I think is the equivalent pipeline using tr as a comment. Then we run the filename through iconv to deal with character set issues. Finally, we split the name into its path and filename components, since we may have to make a new directory before executing the mv.
Here is an update I offer after chepner's answer to avoid nesting bugs. Reverse the output of find with tac to act on folders content before the folders themselves. This way, there is no need to mkdir anymore:
echo "renaming:"
find . -print0 | tac -s '' | while IFS= read -d '' f ;
do
Odir=$(dirname "$f") # original location
Ofile=$(basename "$f") # original filename
newFile=$Ofile
# remove unwanted characters
newFile=$(echo $newFile | tr -d ",'\"?()[]{}\\!")
newFile="${newFile// /_}" # Replace spaces with _
newFile="${newFile//&/n}" # Replace ampersand with n
newFile="${newFile//#/a}" # Replace at sign with a
newFile=$( iconv -f UTF8 -t ASCII//TRANSLIT <<< "$newFile" )
if [[ "$Ofile" != "$newFile" ]]; then # act if something has changed
echo "$Odir/$Ofile to"
echo "$Odir/$newFile"
mv -i "$Odir/$Ofile" "$Odir/$newFile"
echo ""
fi
done
echo "done."
Enjoy ;)
I am using Linux and would like to replace all files containing the string 000000 with an existing file /home/user/offblack.png but keep the existing filename. I've been working at this for a while with various combinations of -exec and xargs but no luck. So far I have:
find | grep 000000
Which does list all the files I want to change fine. How do I copy and replace these files with my existing offblack.png file?
Here's what I would use:
find (your find args here) \
| xargs fgrep '000000' /dev/null \
| awk -F: '{print $1}' \
| xargs -n 1 -I ORIGINAL_FILENAME /bin/echo /bin/cp /path/to/offblack.png ORIGINAL_FILENAME
Expanding, find all the files you're interested in, grep inside of them for the string '000000' (adding /dev/null to the list of files in case one of the generated fgreps ended up with only one filename - it ensures the output is always formatted as "filename: <line containing '000000'>"), strip out only the filenames, then one-by-one, copy in offblack.png over those files. Note that I inserted a /bin/echo in there. That's your dry-run. Remove the echo to get it to run for real.
If what you mean is that the filenames contain "000000":
find . -type f -a -name '*000000*' -exec /bin/echo /bin/cp /path/to/offblack.png {} \;
Much simpler. :-) Find every file under the current directory with a name containing your string and exec the copy of offblack.png over it. Again, what I've given you there is a dry-run. Remove the echo for your live fire drill. :-)
find . -type f | grep 000000 | tr \\n \\0 | xargs -0i+ cp ~/offblack.png "+"
Let's try and use Bash a bit more:
for read -r filename
do
hit=""
for read -r
do
if [[ $REPLY == *000000* ]]
then
hit=$filename
break
fi
done < $filename
[[ -n $hit ]] && cp /path/offblack.png $filename
done < <(find . -type -f)
Fewer man pages to search!
What is the good way to add file extension ".jpg" to extension-less files with bash?
# Strip .jpg from all filenames
for f in *.jpg; do mv "$f" "${f%.jpg}"; done
# Add .jpg to all filenames (even those with .jpg already)
for f in *; do mv "$f" "$f.jpg"; done
# Add .jpg to all filenames...unless they are already .jpg
for f in *; do case "$f" in *.jpg) echo skipped $f;; *) mv "$f" "$f".jpg; esac; done
# Add .jpg to all filenames...unless they already have a . extension
for f in *; do case "$f" in *.*) echo skipped $f;; *) mv "$f" "$f".jpg; esac; done
You can use rename:
rename 's/(.*)/$1.jpg/' *
Another way - without loops
find . -type f -not -name "*.*" -print0 |\
xargs -0 file |\
grep 'JPEG image data' |\
sed 's/:.*//' |\
xargs -I % echo mv % %.jpg
Breakdown:
find all files without extension
check the file type
filter out only JPG files
delete filetype info
xargs run the "mv" for each file
the above command is for dry run, after it you should remove the "echo" before mv
EDIT
Some people suggesting that here is needed "Wrap path arguments in quotes; avoids argument splitting on paths with spaces".
Usually, this recommendation is true, in this case isn't. Because, here the % is got replaced not by shell expansion but by the xargs internally (directly), so the % will be substituted correctly even with spaces in filenames.
Simple demo:
$ mkdir xargstest
$ cd xargstest
# create two files with spaces in names
$ touch 'a b' 'c d'
$ find . -type f -print
./c d
./a b
# notice, here are spaces in the above paths
#the actual xargs mv WITHOUT quotes
$ find . -type f -print | xargs -I % mv % %.ext
$ find . -type f -print
./a b.ext
./c d.ext
# the result is correct even in case with spaces in the filenames...
Simple,
cd to the directory where your files are and:
for f in *;do mv $f $f.jpg;done
dry run:
rename -n s/$/.jpg/ *
actual renaming:
rename s/$/.jpg/ *
find . | while read FILE; do if [ $(file --mime-type -b "$FILE") == "image/jpeg" ]; then mv "$FILE" "$FILE".jpg; fi; done;
In my case i was not aware of the filetype so i used the mv command with the help of the file command to examine and possibly find the file type. This solution might not be perfect for all files since the file command might not recognize the filetype but it worked mostly good for me.
for f in *; do ext=$(file $f | awk '{print $2;}'); mv -n "$f" "$f.$ext"; done
The use of awk is to strip the second word of the string returned from the command file that is actually the extension.
rename --dry-run * -a ".jpg" # test
* -a ".jpg" # rename
You can use move multiple files. I am a maintainer of this project. The syntax is simple.
mmf files*
It will open your $EDITOR with all files names, or vim by default and you can simply highlight the end of all file names using Ctrl+v+G in vim , save the file,quit and that it , all your files are renamed
Ryan Li
The correct syntax for adding a file extension to multiple files within a directory which do not have a file extension is
find . | while read FILE; do if [[ -n `file --mime-type "$FILE" | grep 'message/rfc822'` ]]; then mv "$FILE" "$FILE".eml; fi; done;
I have a number of files in a folder, and I want to replace every space character in all file names with underscores. How can I achieve this?
This should do it:
for file in *; do mv "$file" `echo $file | tr ' ' '_'` ; done
I prefer to use the command 'rename', which takes Perl-style regexes:
rename "s/ /_/g" *
You can do a dry run with the -n flag:
rename -n "s/ /_/g" *
Use sh...
for i in *' '*; do mv "$i" `echo $i | sed -e 's/ /_/g'`; done
If you want to try this out before pulling the trigger just change mv to echo mv.
If you use bash:
for file in *; do mv "$file" ${file// /_}; done
What if you want to apply the replace task recursively? How would you do that?
Well, I just found the answer myself. Not the most elegant solution, (also tries to rename files that do not comply with the condition) but it works. (BTW, in my case I needed to rename the files with '%20', not with an underscore)
#!/bin/bash
find . -type d | while read N
do
(
cd "$N"
if test "$?" = "0"
then
for file in *; do mv "$file" ${file// /%20}; done
fi
)
done
Here is another solution:
ls | awk '{printf("\"%s\"\n", $0)}' | sed 'p; s/\ /_/g' | xargs -n2 mv
uses awk to add quotes around the name of the file
uses sed to replace space with underscores; prints the original name with quotes(from awk); then the substituted name
xargs takes 2 lines at a time and passes it to mv
Try something like this, assuming all of your files were .txt's:
for files in *.txt; do mv “$files” `echo $files | tr ‘ ‘ ‘_’`; done
Quote your variables:
for file in *; do echo mv "'$file'" "${file// /_}"; done
Remove the "echo" to do the actual rename.
To rename all the files with a .py extension use,
find . -iname "*.py" -type f | xargs -I% rename "s/ /_/g" "%"
Sample output,
$ find . -iname "*.py" -type f
./Sample File.py
./Sample/Sample File.py
$ find . -iname "*.py" -type f | xargs -I% rename "s/ /_/g" "%"
$ find . -iname "*.py" -type f
./Sample/Sample_File.py
./Sample_File.py
This will replace ' ' with '_' in every folder and file name recursivelly in Linux with Python >= 3.5. Change path_to_your_folder with your path.
Only list files and folders:
python -c "import glob;[print(x) for x in glob.glob('path_to_your_folder/**', recursive=True)]"
Replace ' ' with '_' in every folder and file name
python -c "import os;import glob;[os.rename(x,x.replace(' ','_')) for x in glob.glob('path_to_your_folder/**', recursive=True)]"
With Python < 3.5, you can install glob2
pip install glob2
python -c "import os;import glob2;[os.rename(x,x.replace(' ','_')) for x in glob2.glob('path_to_your_folder/**')]"
The easiest way to replace a string (space character in your case) with another string in Linux is using sed. You can do it as follows
sed -i 's/\s/_/g' *
Hope this helps.