How to include system variable in batch file inside FOR loop ? also escaping the ! and % sign inside variable? - windows

lets say i have a list file contains folder names that i want to delete periodically based on a list,
currently using this batch file which don't work as i expected :
setlocal enabledelayedexpansion
for /F "tokens=2* delims==" %%x in ('findstr/brc:"foldertodelete" garbagefolderlist.txt') do (
set "foldertodelete=%%x"
set foldertodelete=!foldertodelete:"=!
set foldertodelete=!foldertodelete:%%=^!!"
echo !foldertodelete!
if exist !foldertodelete! (
echo deleting !foldertodelete!
rmdir /s /q !foldertodelete!>nul
)
)
endlocal
inside garbagefolderlist.txt :
foldertodelete="%programfiles%\blablabla"
fodlertodelete=%systemroot%\blablabla
foldertodelete="C:\Temp"
foldertodelete=D:\Temporary files\here
notes about the list file (garbagefolderlist.txt) :
1. folder names may contains double quotes or not, so i want to dynamically eliminate the double quotes inside batch file
2. folder names may be plain or using system variable or not like %systemroot%, etc
3. folder names may contains spaces

If all you want to do is delete the folders listed in that text file you only need one single line of code. You need to use the CALL command to get the double variable expansion that you require. You don't even need delayed expansion at all.
for /F "tokens=2* delims==" %%x in ('findstr /bc:"foldertodelete" garbagefolderlist.txt') do call rmdir /s /q "%%~x" 2>nul
Here is the execution of your script on my system. I created a folder in Program Files and I also have a Temp folder on the C: drive. The line in your file with %systemroot% will not be chosen because you have a typo on that line. So the script will only attempt to process three lines from your input example.
I have added the echo to the code and removed the error dump to nul so that you can see all the output.
#echo off
for /F "tokens=2* delims==" %%x in ('findstr /bc:"foldertodelete" garbagefolderlist.txt') do (
call echo %%~x
call rmdir /s /q "%%~x"
)
And here is the output of that code.
C:\BatchFiles\SO\71120676>so.bat
C:\Program Files\blablabla
Access is denied.
C:\Temp
D:\Temporary files\here
The system cannot find the path specified.
So lets break down that output.
You can see that the %programfiles% variable is expanded as it echo's the folder name correctly but the folder cannot be deleted because I am not running from an elevated cmd prompt as Administrator. So that folder cannot be deleted.
The temp folder displays correctly and is deleted.
The last directory does not exist on my system as I don't have a D: drive so the system reports that it cannot find the path. If standard error was still being redirected to the NUL device you would not see the error which is why I don't bother with checking if a folder exists before I delete it.

Related

How do you loop inside sub directories in windows, and repeat command using windows batch scripting

I am on Windows 10. I have a folder with sub folder. All subfolders have "Subs" folder inside them. I want to loop through all subdirectories, goto Subs directory, unrar a file, Goto next subdirectory and repeat.
I tried below script but could not get to work.
#echo off
setlocal enableextensions enabledelayedexpansion
set "rootFolder=C:\Users\MNM\MAT"
echo ----------------------------------------------------------------------
for /d /r "%rootFolder%" %%a in (.) do (
set mypath=%cd%
#echo %mypath%
cd %%a
set mypath=%cd%
#echo %mypath%
cd Subs
set mypath=%cd%
#echo %mypath%
C:\Users\MNM\MAT\unrar e *subs.rar C:\Users\MNM\mat2\
cd C:\Users\MNM\MAT
)
This simple task can be done with just a single command line:
#echo off
for /R "%USERPROFILE%\MAT" %%I in ("*subs.rar") do "%USERPROFILE%\MAT\UnRAR.exe" x -c- -idcdp -y "%%I" "%USERPROFILE%\mat2\"
USERPROFILE is a predefined Windows Environment Variable which is on your computer for your user account defined with value C:\Users\MNM.
The command FOR searches in directory C:\Users\MNM\MAT and all its non hidden subdirectories because of /R for non hidden files matching the pattern *subs.rar. Each file name found is assigned with full path to loop variable I.
UnRAR is executed for each found RAR archive file for extracting the archive to directory C:\Users\MNM\mat2 with extracting also the directory structures inside the RAR archive file because of command x instead of e. Existing files in destination directory (tree) are automatically overwritten because of -y. The switches -c- and -idcdp are for displaying less information during extraction process.
For a brief description of used and additionally available switches of UnRAR run in command prompt window UnRAR without any parameter or with /? as parameter. A complete description of the commands and switches of UnRAR can be found in text file Rar.txt in program files folder of WinRAR if that shareware application is also installed and not just the freeware UnRAR.
It is absolutely not needed to change into the directory containing the RAR archive file on extracting all RAR archives into same destination directory as it can be seen here.
This is one possible way if I understood your folder structure correctly:
#echo off
set "Base=C:\Users\MNM\MAT"
echo ----------------------------------------------------------------------
for /F "delims=" %%A in (
'dir /B/S "%Base%\*subs.rar" ^| findstr /i "^%Base:\=\\%\\[^\\]*\\Subs\\[^\\]*subs.rar$"'
) do Echo "C:\Users\MNM\MAT\unrar.exe" e "%%~fA" "C:\Users\MNM\mat2\"
the for /f will parse the output of the dir and findstr
dir will iterate all *subs.rar in the tree starting from %Base%
the complex RegEx in findstr will filter the rar's to those in a folder subs in a subfolder of %Base%
as a backslash is an escape char in a RegEx, literal backslashes have to be doubled.
If the output looks ok remove the echo in the last line.
Just because recursing all subdirectories and extracting all *subs.rar files wasn't requested here's an example that is based upon my assumptions:
#ECHO OFF
SET "rootDir=%USERPROFILE%\MAT"
IF /I NOT "%CD%"=="%rootDir%" CD /D "%rootDir%"
FOR /D %%A IN (*
) DO IF EXIST "%%A\Subs\*subs.rar" UNRAR e "%%A\Subs\*subs.rar" mat2\

How to Create Conditioned Subfolder through batch file?

We are using Jenkins as our CI tool. At present, I have written a batch file, which takes build value from a notepad file (written by developer), and then copies it to the network drive with the build value as a folder.
The batch file is mentioned below
for /f "tokens=2" %%i in ('findstr Build "C:\Program Files (x86)\Jenkins\jobs\**\build.txt"') do set Build=%%i
set Build=%Build:'=%
if not exist "\%Build%\" mkdir "%Build%\"
My motive now is that, if the value in notepad is 6.1 or 6.a, i.e anything after decimal,it doesn't create folder, but places it in 6 subfolder. Alternatively, if the value is full, it creates a parent folder.
The text file, where I get the build number is mentioned below.
#define MyAppVersion '4.0.0.0'
#define MyFullAppVersion '4.1.0.0'
#define BuildNumber '81'
I need to create folder on the basis of value entered in "Build Number". If the value is in decimal, it should create within subfolder, else there should be only a main folder.
With the help of below input, I created below batch file for achieving this, but it didn't worked.
for /f "tokens=2" %%i in ('findstr Build "C:\Folder Check\logs\build.txt"') do set Build=%%i
set Build=%Build:'=%
for /F "tokens=2""delims=." %%j in ('findstr Build "C:\Folder Check\logs\build.txt"') do set "sub=%%j"
if exist "\\network\%Version%\%Build%\" mkdir "\\network\%Build%\%sub%%"
if not exist "\\network\%Version%\%Build%\" mkdir "\\network\%Version%\%Build%\"
if exist "\\network\%Version%\%Build%\%sub%%j" XCOPY /y "C:\Program Files (x86)\Notepad++\readme.txt" "\\network\%Version%\%Build%\%sub%\" /E /S
if not exist "\\172.19.0.4\Departement$\Development\Development RCX\RCX_M02_CSP\INTERNAL RELEASES\CI\%Version%\%Build%\%sub%%" XCOPY /y "C:\Program Files (x86)\Notepad++\readme.txt" "\\network\%Version%\%Build%\" /E /S
Any help will be greatly appreciated.
You could use the following code to split off the period . and everything behind:
for /F "delims=." %%j in ("%Build%") do set "Build=%%j"
Take also a look at the quoted set syntax, which is the one I recommend in general, because no special characters could cause problems that way.
Your if exist syntax appears a bit odd to me; I am not sure where the directory %Build% should be located, because you precede a \ during existence check but you do not during creation. If you want to ensure the directory exists in the current working directory, do not precede the \:
if not exist "%Build%\" mkdir "%Build%"
If you want it to be located in the root directory od the current drive, then do precede it by \, but this is not what you want, I guess.
However, the appended \ is needed to check a directory for existence; not providing it would also check for files.
Anyway, to ensure a directory exists, you do not need to check for existence, you can just create it and suppress the error message that appears in case the directory already exists, like this:
mkdir "%Build%" 2> nul
I am not sure if I interpret the ** in your path specification right, but I assume you are going to let me know...
You can use global wild-cards (*, ?) only in the last element of a path, so C:\Program Files (x86)\Jenkins\jobs\**\build.txt is not valid. In case there is only a single directory in C:\Program Files (x86)\Jenkins\jobs\, you can get it like this:
for /D %%D in ("C:\Program Files (x86)\Jenkins\jobs\*") do set "SubDir=%%~D"
If there are multiple directories and you want to get the last modified one, you could use this:
for /F "eol=: delims=" %%D in ('
dir /B /A:D /O:D "C:\Program Files (x86)\Jenkins\jobs\*"
') do set "SubDir=%%D"

How to execute an application existing in each specific folder of a directory tree on a file in same folder?

I have some folders with different names. Each folder has a specific structure as listed below:
Folder1
Contents
x64
Folder1.aaxplugin
TransVST_Fixer.exe
Folder 2
Contents
x64
Folder 2.aaxplugin
TransVST_Fixer.exe
There are two files within each subfolder x64. One file has the same name as the folder two folder levels above. The other file is an .exe file whose name is the same in all folders.
Now I need to run file with file extension aaxplugin on each specific .exe file. It would be obviously very time consuming opening each and every single folder and drag & drop each file on .exe to run it on this file.
That's why I am trying to create a batch script to save some time.
I looked for solutions here on Stack Overflow. The only thing I have found so far was a user saying this: When I perform a drag & drop, the process 'fileprocessor.exe' is executed. When I try to launch this exe, though, CMD returns error ('not recognized or not batch file' stuff).
How can I do this?
UPDATE 12/22/2015
I used first a batch file with following line to copy the executable into x64 subfolder of Folder1.
for /d %%a in ("C:\Users\Davide\Desktop\test\Folder1\*") do ( copy "C:\Program Files\Sugar Bytes\TransVST\TransVST_Fixer.exe" "%%a\x64\" 2> nul )
After asking here, I tried the following script:
for /f "delims=" %%F in ('dir /b /s x64\*.aaxplugin') do "%%~dpFTransVST_Fixer.exe" "%%F"
Unfortunately, the output is as following
C:\Users\Davide\Desktop>for /F "delims=" %F in ('dir /b /s x64\*.aaxplugin') do "%~dpFTransVST_Fixer.exe" "%F"
The system cannot find the file specified.
Try the following batch code:
#echo off
setlocal EnableDelayedExpansion
for /R "%USERPROFILE%\Desktop\test" %%F in (*.aaxplugin) do (
set "FilePath=%%~dpF"
if not "!FilePath:\x64\=!" == "!FilePath!" "%ProgramFiles%\Sugar Bytes\TransVST\TransVST_Fixer.exe" "%%F"
)
endlocal
The command FOR with option/R searches recursive in all directories of directory %USERPROFILE%\Desktop\test being expanded on your machine to C:\Users\Davide\Desktop for files with file extension aaxplugin. The loop variable F contains on each loop run the name of the found file with full path without surrounding double quotes.
The drive and path of each found file is assigned to environment variable FilePath.
Next a case-sensitive string comparison is done between file path with all occurrences of string \x64\ case-insensitive removed with unmodified file path.
Referencing value of environment variable FilePath must be done here using delayed expansion because being defined and evaluated within a block defined with ( ... ). Otherwise command processor would expand %FilePath% already on parsing the entire block resulting in a syntax error on execution because string substitution is not possible as no environment variable FilePath defined above body block of FOR loop.
The strings are not equal if path of file contains a folder with name x64. This means on provided folder structure that the file is in folder x64 and not somewhere else and therefore the application is executed next from its original location to fix the found *.aaxplugin file.
The line with IF is for the folder structure example:
if not "C:\Users\Davide\Desktop\test\Folder1\Contents" == "C:\Users\Davide\Desktop\test\Folder1\Contents\x64\"
if not "C:\Users\Davide\Desktop\test\Folder 2\Contents" == "C:\Users\Davide\Desktop\test\Folder 2\Contents\x64\"
So for both *.aaxplugin files the condition is true because the compared strings are not identical
Also possible would be:
#echo off
setlocal EnableDelayedExpansion
for /F "delims=" %%F in ('dir /A-D /B /S "%USERPROFILE%\test\*.aaxplugin" 2^>nul') do (
set "FilePath=%%~dpF"
if not "!FilePath:\x64\=!" == "!FilePath!" "%ProgramFiles%\Sugar Bytes\TransVST\TransVST_Fixer.exe" "%%F"
)
endlocal
But command DIR is not really necessary as it can be seen on first provided code.
But if the application TransVST_Fixer.exe for some unknown reason does its job right only with directory of file being also the current directory, the following batch code could be used instead of first code using the commands pushd and popd:
#echo off
setlocal EnableDelayedExpansion
for /R "%USERPROFILE%\test" %%F in (*.aaxplugin) do (
set "FilePath=%%~dpF"
echo !FilePath!
if /I "!FilePath:~-5!" == "\x64\" (
pushd "%%~dpF"
"%ProgramFiles%\Sugar Bytes\TransVST\TransVST_Fixer.exe" "%%~nxF"
popd
)
)
endlocal
There is one more difference in comparison to first code. Now the last 5 characters of path of file are compared case-insensitive with the string \x64\. Therefore the file must be really inside a folder with name x64 or X64. A folder with name x64 or X64 anywhere else in path of file does not result anymore in a true state for the condition as in first two batch codes.
But if for some unknown reason it is really necessary to run the application in same folder as the found *.aaxplugin and the directory of the file must be the current directory, the following batch code could be used:
#echo off
setlocal EnableDelayedExpansion
for /R "%USERPROFILE%\test" %%# in (*.aaxplugin) do (
set "FilePath=%%~dp#"
if /I "!FilePath:~-5!" == "\x64\" (
pushd "%%~dp#"
"%%~dp#TransVST_Fixer.exe" "%%~nx#"
popd
)
)
endlocal
The path of the file referenced with %%~dpF always ends with a backslash which is the reason why there is no backslash left of TransVST_Fixer.exe (although command processor could handle also file with with two backslashes in path).
In batch code above character # is used as loop variable because %%~dp#TransVST_Fixer.exe is easier to read in comparison to %%~dpFTransVST_Fixer.exe. It is more clear for a human with using # as loop variable where the reference to loop variable ends and where name of application begins. For the command processor it would not make a difference if loop variable is # or upper case F.
A lower case f would work here also as loop variable, but is in general problematic as explained on Modify variable within loop of batch script.
For understanding the used commands and how they work, open a command prompt window, execute there the following commands, and read entirely all help pages displayed for each command very carefully.
dir /?
echo /?
endlocal /?
for /?
if /?
popd /?
pushd /?
set /?
setlocal /?
Your question isn't quite clear, but it seems, something like this should work:
for /f "delims=" %%f in ('dir /b /s X64\*.ext') do "%%~dpfMyExe.exe" "%%f"
Maybe you have to change directory to each folder (depends on your .exe):
for /f "delims=" %%d in ('dir /B /ad') do (
pushd "%%d"
for /f "delims=" %%f in ('dir /b "contents\x64\*.ext"') do (
cd Contents\x64
MyExe.exe "%%f"
)
popd
)
Assuming:
The Directory structure is fixed and the files are indeed in a subfolder contents\X64\.
MyExe.exe is the same (name) in every folder.
There is only one file *.ext in every folder.
I'll give you the script I created for doing so, hope it works for you
for /d %%d IN (./*) do (cd "%%d/Contents/x64" & "../../../TransVST_Fixer.exe" "%%d" & cd "/Program Files (x86)\Common Files\Avid\Audio\Plug-Ins")
Please note that I placed the fixer inside the root folder so I just have to copy it once. You have to place it inside your root folder and execute it. What it does:
iterate over each folder
for each one it enters /Contents/x64, executes the fixer (wich is 3 levels above) and after that returns to the original folder.
If you have your plugins in a different folder, you just have to change this part replacing the path for the one you have your plugins in.
cd "/Program Files (x86)\Common Files\Avid\Audio\Plug-Ins"
REMEMBER to place the script on that folder. For this example I place my script on the folder "/Program Files (x86)\Common Files\Avid\Audio\Plug-Ins" and run it (as a .bat).
PS: the fixer will place the fixed plugins in "C:\Users\Public\modified" (just read the screen while executing, it gives you the new files path. If you want to move them to the right path, you can execute this from the new files path ("C:\Users\Public\modified")
for %%d IN (*.aaxplugin) do (mkdir "%%d_temp/Contents\x64" & move "%%d" "%%d_temp/Contents\x64/%%d" & rename "%%d_temp" "%%d")
with that, I iterate over every plugin and create a folder with the same name (I create _temp because of name colision, after moving the file I rename it to the correct one), also with the subfolder "/Contents/x64", and move the plugin inside. Once donde, you can just take the resulting folders and place them in their correct path.
Hope it works, for me it works like a charm.

For loop in batch file reading a file of File Paths

I want to write a Windows batch file script that will loop through a text file of FILE PATHS, do some work using data from each file path, then ultimately delete the file.
I started by running the FORFILES command and sending its output (the #PATH parameter is the full path of any file it matches) to a text file (results.txt).
I end up with a results.txt file like this:
"C:/Windows/Dir1/fileA.log"
"C:/Windows/Dir1/fileA.log"
"C:/Windows/Dir2/fileC.log"
"C:/Windows/Dir3/fileB.log"
What I want to do is:
Use a FOR loop and read each line in the results.txt file
For each line (file path), strip out the directory name that the log file is sitting in (ie: Dir1, Dir2, etc..) and create a directory with that SAME name in a different location (ie. D:/Archive/Backups/Dir1, D:/Archive/Backups/Dir2, etc..) -- assuming the directory doesn't exist.
Move the actual .log file to a zip file in that directory [I have code to do this].
Delete the .log file from its original location. [Pretty straightforward]
I'm having trouble figuring out the best way to accomplish the first 2 steps. My FOR loop seems to stop after reading the very first line:
FOR /F "tokens=1,2,3,4,5,6,7,8,9,10 delims=\" %%G in ("results.txt") DO (
...
)
You don't want to parse the path with the tokens/delims options because you don't know how many directory levels you are dealing with. You want to preserve each line in its entirety. TO parse the path you want to use the FOR variable modifiers. (type HELP FOR from the command line and look at the last section of the output)
%%~pG gives the path (without the drive or file name). If we then strip off the last \, we can go through another FOR iteration and get the name (and possible extension) of the file's directory by using %%~nxA.
The toggling of delayed expansion is just to protect against a possible ! in the path. If you know that no path contains ! then you can simply enable delayed expansion at the top of the script and be done with it.
EDIT - this code has been modified significantly since Aacini pointed out that I misread the requirements. It should satisfy the requirements now.
for /f "usebackq delims=" %%G in ("results.txt") do (
set "myPath=%~pG"
setlocal enableDelayedExpansion
for /f "eol=: delims=" %%A in ("!myPath:~0,-1!") do (
endlocal
if not exist d:\Archive\Backups\%%~nxA md d:\Archive\Backups\%%~nxA
rem ::zip %%G into zip file in the D: location
rem ::you should be able to create the zip with the move option
rem ::so you don't have to del the file
)
)
I wrote this to timestamp files before offloading to SFTP.
Hope you find it useful.
The timestamp coding may seem irrelevant to your issue, but I left it because it's a good example of dissecting the filename itself.
I suggest you put an ECHO in front of the REN command for testing. Different shells may have different results.
In the end, the delayedexpansion command wasn't necessary. It was the sub-routine that fixed my issues with variables inside the loop. That could possibly be because of my OS ver. (Win 8.1) - It wouldn't hurt to leave it.
#echo off
cls
setlocal enabledelayedexpansion
if %time:~0,2% geq 10 set TIMESTAMP=%date:~10,4%%date:~4,2%%date:~7,2%_%time:~0,2%%time:~3,2%%time:~6,2%
if %time:~0,2% leq 9 set TIMESTAMP=%date:~10,4%%date:~4,2%%date:~7,2%_0%time:~1,1%%time:~3,2%%time:~6,2%
echo TimeStamp=%TIMESTAMP%
echo.
for %%G in (*.txt) do (
set OLDNAME=%%G
call :MXYZPTLK
)
dir *.txt
goto :EOF
:MXYZPTLK
echo OldName=%OLDNAME%
ren %OLDNAME% %OLDNAME:~0,-4%_%TIMESTAMP%%OLDNAME:~-4,4%
echo.
:END
You have two minor problems:
The path separator in the file is '/' but you use '\' in the for loop.
The quotes around "results.txt" stop it working.
This works. Don't write quotes to results.txt and you won't get a quote at the end of the filename.
#echo off
FOR /F "tokens=3,4 delims=/" %%I in (results.txt) DO (
REM Directory
echo %%I
REM File
echo %%J
)

How to delete files/subfolders in a specific directory at the command prompt in Windows

Say, there is a variable called %pathtofolder%, as it makes it clear it is a full path of a folder.
I want to delete every single file and subfolder in this directory, but not the directory itself.
But, there might be an error like 'this file/folder is already in use'... when that happens, it should just continue and skip that file/folder.
Is there some command for this?
rmdir is my all time favorite command for the job. It works for deleting huge files and folders with subfolders. A backup is not created, so make sure that you have copied your files safely before running this command.
RMDIR "FOLDERNAME" /S /Q
This silently removes the folder and all files and subfolders.
You can use this shell script to clean up the folder and files within C:\Temp source:
del /q "C:\Temp\*"
FOR /D %%p IN ("C:\Temp\*.*") DO rmdir "%%p" /s /q
Create a batch file (say, delete.bat) containing the above command. Go to the location where the delete.bat file is located and then run the command: delete.bat
The simplest solution I can think of is removing the whole directory with
RD /S /Q folderPath
Then creating this directory again:
MD folderPath
This will remove the folders and files and leave the folder behind.
pushd "%pathtofolder%" && (rd /s /q "%pathtofolder%" 2>nul & popd)
#ECHO OFF
SET THEDIR=path-to-folder
Echo Deleting all files from %THEDIR%
DEL "%THEDIR%\*" /F /Q /A
Echo Deleting all folders from %THEDIR%
FOR /F "eol=| delims=" %%I in ('dir "%THEDIR%\*" /AD /B 2^>nul') do rd /Q /S "%THEDIR%\%%I"
#ECHO Folder deleted.
EXIT
...deletes all files and folders underneath the given directory, but not the directory itself.
CD [Your_Folder]
RMDIR /S /Q .
You'll get an error message, tells you that the RMDIR command can't access the current folder, thus it can't delete it.
Update:
From this useful comment (thanks to Moritz Both), you may add && between, so RMDIR won't run if the CD command fails (e.g. mistyped directory name):
CD [Your_Folder] && RMDIR /S /Q .
From Windows Command-Line Reference:
/S: Deletes a directory tree (the specified directory and all its
subdirectories, including all files).
/Q: Specifies quiet mode. Does not prompt for confirmation when
deleting a directory tree. (Note that /q works only if /s is
specified.)
I use Powershell
Remove-Item c:\scripts\* -recurse
It will remove the contents of the folder, not the folder itself.
RD stands for REMOVE Directory.
/S : Delete all files and subfolders
in addition to the folder itself.
Use this to remove an entire folder tree.
/Q : Quiet - do not display YN confirmation
Example :
RD /S /Q C:/folder_path/here
Use Notepad to create a text document and copy/paste this:
rmdir /s/q "%temp%"
mkdir "%temp%"
Select Save As and file name:
delete_temp.bat
Save as type: All files and click the Save button.
It works on any kind of account (administrator or a standard user). Just run it!
I use a temporary variable in this example, but you can use any other! PS: For Windows OS only!
None of the answers as posted on 2018-06-01, with the exception of the single command line posted by foxidrive, really deletes all files and all folders/directories in %PathToFolder%. That's the reason for posting one more answer with a very simple single command line to delete all files and subfolders of a folder as well as a batch file with a more complex solution explaining why all other answers as posted on 2018-06-01 using DEL and FOR with RD failed to clean up a folder completely.
The simple single command line solution which of course can be also used in a batch file:
pushd "%PathToFolder%" 2>nul && ( rd /Q /S "%PathToFolder%" 2>nul & popd )
This command line contains three commands executed one after the other.
The first command PUSHD pushes current directory path on stack and next makes %PathToFolder% the current directory for running command process.
This works also for UNC paths by default because of command extensions are enabled by default and in this case PUSHD creates a temporary drive letter that points to that specified network resource and then changes the current drive and directory, using the newly defined drive letter.
PUSHD outputs following error message to handle STDERR if the specified directory does not exist at all:
The system cannot find the path specified.
This error message is suppressed by redirecting it with 2>nul to device NUL.
The next command RD is executed only if changing current directory for current command process to specified directory was successful, i.e. the specified directory exists at all.
The command RD with the options /Q and /S removes a directory quietly with all subdirectories even if the specified directory contains files or folders with hidden attribute or with read-only attribute set. The system attribute does never prevent deletion of a file or folder.
Not deleted are:
Folders used as the current directory for any running process. The entire folder tree to such a folder cannot be deleted if a folder is used as the current directory for any running process.
Files currently opened by any running process with file access permissions set on file open to prevent deletion of the file while opened by the running application/process. Such an opened file prevents also the deletion of entire folder tree to the opened file.
Files/folders on which the current user has not the required (NTFS) permissions to delete the file/folder which prevents also the deletion of the folder tree to this file/folder.
The first reason for not deleting a folder is used by this command line to delete all files and subfolders of the specified folder, but not the folder itself. The folder is made temporarily the current directory for running command process which prevents the deletion of the folder itself. Of course this results in output of an error message by command RD:
The process cannot access the file because it is being used by another process.
File is the wrong term here as in reality the folder is being used by another process, the current command process which executed command RD. Well, in reality a folder is for the file system a special file with file attribute directory which explains this error message. But I don't want to go too deep into file system management.
This error message, like all other error messages, which could occur because of the three reasons written above, is suppressed by redirecting it with 2>nul from handle STDERR to device NUL.
The third command, POPD, is executed independently of the exit value of command RD.
POPD pops the directory path pushed by PUSHD from the stack and changes the current directory for running the command process to this directory, i.e. restores the initial current directory. POPD deletes the temporary drive letter created by PUSHD in case of a UNC folder path.
Note: POPD can silently fail to restore the initial current directory in case of the initial current directory was a subdirectory of the directory to clean which does not exist anymore. In this special case %PathToFolder% remains the current directory. So it is advisable to run the command line above not from a subdirectory of %PathToFolder%.
One more interesting fact:
I tried the command line also using a UNC path by sharing local directory C:\Temp with share name Temp and using UNC path \\%COMPUTERNAME%\Temp\CleanTest assigned to environment variable PathToFolder on Windows 7. If the current directory on running the command line is a subdirectory of a shared local folder accessed using UNC path, i.e. C:\Temp\CleanTest\Subfolder1, Subfolder1 is deleted by RD, and next POPD fails silently in making C:\Temp\CleanTest\Subfolder1 again the current directory resulting in Z:\CleanTest remaining as the current directory for the running command process. So in this very, very special case the temporary drive letter remains until the current directory is changed for example with cd /D %SystemRoot% to a local directory really existing. Unfortunately POPD does not exit with a value greater 0 if it fails to restore the initial current directory making it impossible to detect this very special error condition using just the exit code of POPD. However, it can be supposed that nobody ever runs into this very special error case as UNC paths are usually not used for accessing local files and folders.
For understanding the used commands even better, open a command prompt window, execute there the following commands, and read the help displayed for each command very carefully.
pushd /?
popd /?
rd /?
Single line with multiple commands using Windows batch file explains the operators && and & used here.
Next let us look on the batch file solution using the command DEL to delete files in %PathToFolder% and FOR and RD to delete the subfolders in %PathToFolder%.
#echo off
setlocal EnableExtensions DisableDelayedExpansion
rem Clean the folder for temporary files if environment variable
rem PathToFolder is not defined already outside this batch file.
if not defined PathToFolder set "PathToFolder=%TEMP%"
rem Remove all double quotes from folder path.
set "PathToFolder=%PathToFolder:"=%"
rem Did the folder path consist only of double quotes?
if not defined PathToFolder goto EndCleanFolder
rem Remove a backslash at end of folder path.
if "%PathToFolder:~-1%" == "\" set "PathToFolder=%PathToFolder:~0,-1%"
rem Did the folder path consist only of a backslash (with one or more double quotes)?
if not defined PathToFolder goto EndCleanFolder
rem Delete all files in specified folder including files with hidden
rem or read-only attribute set, except the files currently opened by
rem a running process which prevents deletion of the file while being
rem opened by the application, or on which the current user has not
rem the required permissions to delete the file.
del /A /F /Q "%PathToFolder%\*" >nul 2>nul
rem Delete all subfolders in specified folder including those with hidden
rem attribute set recursive with all files and subfolders, except folders
rem being the current directory of any running process which prevents the
rem deletion of the folder and all folders above, folders containing a file
rem opened by the application which prevents deletion of the file and the
rem entire folder structure to this file, or on which the current user has
rem not the required permissions to delete a folder or file in folder tree
rem to delete.
for /F "eol=| delims=" %%I in ('dir "%PathToFolder%\*" /AD /B 2^>nul') do rd /Q /S "%PathToFolder%\%%I" 2>nul
:EndCleanFolder
endlocal
The batch file first makes sure that environment variable PathToFolder is really defined with a folder path without double quotes and without a backslash at the end. The backslash at the end would not be a problem, but double quotes in a folder path could be problematic because of the value of PathToFolder is concatenated with other strings during batch file execution.
Important are the two lines:
del /A /F /Q "%PathToFolder%\*" >nul 2>nul
for /F "eol=| delims=" %%I in ('dir "%PathToFolder%\*" /AD /B 2^>nul') do rd /Q /S "%PathToFolder%\%%I" 2>nul
The command DEL is used to delete all files in the specified directory.
The option /A is necessary to process really all files including files with the hidden attribute which DEL would ignore without using option /A.
The option /F is necessary to force deletion of files with the read-only attribute set.
The option /Q is necessary to run a quiet deletion of multiple files without prompting the user if multiple files should be really deleted.
>nul is necessary to redirect the output of the file names written to handle STDOUT to device NUL of which can't be deleted because of a file is currently opened or user has no permission to delete the file.
2>nul is necessary to redirect the error message output for each file which can't be deleted from handle STDERR to device NUL.
The commands FOR and RD are used to remove all subdirectories in specified directory. But for /D is not used because of FOR is ignoring in this case subdirectories with the hidden attribute set. For that reason for /F is used to run the following command line in a separate command process started in the background with %ComSpec% /c:
dir "%PathToFolder%\*" /AD /B 2>nul
DIR outputs in bare format because of /B the directory entries with attribute D, i.e. the names of all subdirectories in specified directory independent on other attributes like the hidden attribute without a path. 2>nul is used to redirect the error message output by DIR on no directory found from handle STDERR to device NUL.
The redirection operator > must be escaped with the caret character, ^, on the FOR command line to be interpreted as a literal character when the Windows command interpreter processes this command line before executing the command FOR which executes the embedded dir command line in a separate command process started in the background.
FOR processes the captured output written to handle STDOUT of a started command process which are the names of the subdirectories without path and never enclosed in double quotes.
FOR with option /F ignores empty lines which don't occur here as DIR with option /B does not output empty lines.
FOR would also ignore lines starting with a semicolon which is the default end of line character. A directory name can start with a semicolon. For that reason eol=| is used to define the vertical bar character as the end-of-line character which no directory or file can have in its name.
FOR would split up the line into substrings using space and horizontal tab as delimiters and would assign only the first space/tab delimited string to specified loop variable I. This splitting behavior is not wanted here because of a directory name can contain one or more spaces. Therefore delims= is used to define an empty list of delimiters to disable the line splitting behavior and get assigned to the loop variable, I, always the complete directory name.
Command FOR runs the command RD for each directory name without a path which is the reason why on the RD command line the folder path must be specified once again which is concatenated with the subfolder name.
For understanding the used commands and how they work, open a command prompt window, execute there the following commands, and read entirely all help pages displayed for each command very carefully.
del /?
dir /?
echo /?
endlocal /?
for /?
goto /?
if /?
rd /?
rem /?
set /?
setlocal /?
To delete file:
del PATH_TO_FILE
To delete folder with all files in it:
rmdir /s /q PATH_TO_FOLDER
To delete all files from specific folder (not deleting folder itself) is a little bit complicated. del /s *.* cannot delete folders, but removes files from all subfolder. So two commands are needed:
del /q PATH_TO_FOLDER\*.*
for /d %i in (PATH_TO_FOLDER\*.*) do #rmdir /s /q "%i"
You can do it by using the following command to delete all contents and the parent folder itself:
RMDIR [/S] [/Q] [drive:]path
#ECHO OFF
rem next line removes all files in temp folder
DEL /A /F /Q /S "%temp%\*.*"
rem next line cleans up the folder's content
FOR /D %%p IN ("%temp%\*.*") DO RD "%%p" /S /Q
I tried several of these approaches, but none worked properly.
I found this two-step approach on the site Windows Command Line:
forfiles /P %pathtofolder% /M * /C "cmd /c if #isdir==FALSE del #file"
forfiles /P %pathtofolder% /M * /C "cmd /c if #isdir==TRUE rmdir /S /Q #file"
It worked exactly as I needed and as specified by the OP.
I had following solution that worked for me:
for /R /D %A in (*node_modules*) do rmdir "%A" /S /Q
It removes all node modules folder from current directory and its sub-folders.
This is similar to solutions posted above, but i am still posting this here, just in case someone finds it useful
Use:
del %pathtofolder%\*.* /s /f /q
This deletes all files and subfolders in %pathtofolder%, including read-only files, and does not prompt for confirmation.

Resources