Batch script with for loop and pipe - windows

I would like all the csv files in a directory which filename does not contain word "summary". Inside the command prompt I can type the following command
dir /b my_dir\*.csv | find /V "summary"
When I try to transfer the above command into a batch file I run into a problem in that the pipe command is not supported in the for loop. That is I cannot do the following
FOR /f %%A in ('dir /b my_dir\*.csv | find /V "summary"') do (
rem want to do something here
)
Can somebody shed some light to me on how to solve the problem above?
Thanks in advance!

You need to escape the | character to prevent its being interpreted at the time of parsing the loop command. Use ^ to escape it:
FOR /f "delims=" %%A in ('dir /b "my_dir\*.csv" ^| find /V "summary"') do (
rem do what you want with %%A here
)
Once escaped, the | becomes part of the '-delimited string. It is only interpreted as a special symbol when that string is parsed separately from the loop, as a "sub-command", according to the syntax. And that is done after parsing the loop.

If you get the problem that Gilbeg got "find: /V': No such file or directory" then it's most likely you have cygwin, or similar, in your path and the batch file's not using the Windows find command. If you modify your script to use the absolute path of the Windows find then the error will go away:
FOR /f "delims=" %%A in ('dir /b "my_dir\*.csv" ^| %SYSTEMROOT%\system32\find.exe /V "summary"') do (
rem want to do something here with %%A
)

You can also just embed a double-quoted string inside the single-quotes string, as in:
FOR /f "delims=" %%A in ('"dir /b my_dir\*.csv | find /I /V "summary""') do #(
ECHO Do something with "%%A"
)

have a look at the Windows PowerShell. Not that I have ever used it myself, mind.

Related

Quoting a long filenamed command in a for loop in a batch file

Linked:
Best free resource for learning advanced batch-file usage?
Dealing with quotes in Windows batch scripts
This appears to be one of those maddening quoting issues. In this example program:
#echo off
set wmicpath=%windir%\System32\wbem\wmic.exe
for /f "usebackq" %%a in (`%wmicpath% COMPUTERSYSTEM GET SystemType ^| findstr /I "x64"`) do (
echo %%a
)
The program runs just fine. Unless you try to quote the wmicpath. Imagine if you will that it contains a long path name. Then you should quote it. But I cannot quite get it to work. This fails:
for /f "usebackq" %%a in (`"%wmicpath%" COMPUTERSYSTEM GET SystemType ^| findstr /I "x64"`) do (
but this works!:
for /f "usebackq" %%a in (`"%wmicpath%" COMPUTERSYSTEM GET SystemType ^| findstr /I x64`) do (
as does this:
for /f "usebackq" %%a in (`"%wmicpath%" COMPUTERSYSTEM GET SystemType`) do (
There's something really odd about matching quotes in a for command. You can quote a command as long as you don't start quoting elsewhere...
Is it possible? I tried escaping at various points but I'm not sure about the escaping rules when quotes are involved...
Edit: I think this link might be the issue (ie: it's a bug): Pipe in for loop breaks double quoted variables
#echo off
setlocal enableextensions disabledelayedexpansion
set "wmicpath=%windir%\System32\wbem\wmic.exe"
for /f "usebackq delims=" %%a in (`
^""%wmicpath%" COMPUTERSYSTEM GET SystemType ^| findstr /I "x64"^"
`) do (
echo %%a
)
If you look at the start and end of the inner command, you will see two additional ^" (a escaped double quote). Your problem is that the for command is spawning a separate instance of cmd to handle the inner command, and this separate instance is removing the initial and final double quotes.
Why escaped quotes? To avoid this additional quotes being paired with the double quotes in the command that could lead to some other parsing problems.
You can run cmd /? to obtain the help page (sorry, i have a spanish locale so i will not include the output here). You will see a section about the /C and /K usage explaining quote removal behaviour.
First of all I would change the command, WMIC allows you to use a query language LIKE operator which would in this case remove the need to pipe anything.
#Echo Off
Set "WMIC=%SystemRoot%\System32\Wbem\wmic.exe"
For /F "UseBackQ Skip=1" %%a In (
`""%WMIC%" ComputerSystem Where "SystemType Like 'x64%%'" Get SystemType"`
) Do For %%b In (%%a) Do Echo=%%b
Timeout -1
Then I may even change the format of the command such that I don't use back quotes.
#Echo Off
Set "WMIC=%SystemRoot%\System32\Wbem\wmic.exe"
For /F "Skip=1" %%a In (
'""%WMIC%" ComputerSystem Where (SystemType Like "x64%%") Get SystemType"'
) Do For %%b In (%%a) Do Echo=%%b
Timeout -1
Whilst this doesn't directly answer the question in the subject title, it does allow for your particular command to work correctly.
However neither are necessary to your particular command example, because you do not need the for loop to echo that output:
#Echo Off
Set "WMIC=%SystemRoot%\System32\Wbem\wmic.exe"
"%WMIC%" ComputerSystem Get SystemType|Find /I "x64"
Timeout -1
Change Find to FindStr if you feel the need.
>x64.txt ECHO x64
for /f "usebackq" %%a in (`"%wmicpath%" COMPUTERSYSTEM GET SystemType ^| findstr /I /g:x64.txt`) do (
might be a work-around, depending on your actual application and preferences.

Truncate filename after a multiple of a special character in windows batch script?

I want to remove the part of a filename after the third "_" from thousand of files. The structure after the third "_" varies and contains "_" in some cases. The length of the first part varies so I can't just remove the first 15 characters. The result should be unique.
The filenames look like this:
00_TEXT_=Text00._AA1234L_AA1_1.pdf
00_TEX_=Text00._AA1234L_AA1_2.pdf
00_TEXT_=TextText00._DD2023A.pdf
00_TEXT_=Text00._AA2345L_BB1_1.pdf
00_TEXT_=Text00._AA2345L_BB1_2.pdf
The result should look like this:
AA1234L_AA1_1.pdf
AA1234L_AA1_2.pdf
DD2023A.pdf
AA2345L_BB1_1.pdf
AA2345L_BB1_2.pdf
Any idea why this is not working:
#echo off
setlocal enabledelayedexpansion
set deletestring=*_*_*_
for /f "delims==" %%F in ('dir /b ^| find "%deletestring%"') do (
set oldfilename=%%F
set newfilename=!oldfilename:%deletestring%=!
Ren "!oldfilename!" "!newfilename!"
)
I was able to get it working with this:
#echo off
setlocal enabledelayedexpansion
set deletestring=*_*_*_*
for /f "tokens=1,2,3,* delims=_" %%F in ('dir /b "%deletestring%"') do (
Ren "%%F_%%G_%%H_%%I" "%%I"
)
endlocal
Note that enabledelayedexpansion isn't really needed in the above.
Alternately, you could do this as a single line (no batch file needed):
for /f "tokens=1,2,3,* delims=_" %F in ('dir /b "*_*_*_*"') do Ren "%F_%G_%H_%I" "%I"
The idea is to simply split the matching filenames apart by underscores and then reconstruct the names during the rename process (%%F_%%G_%%H_%%I gives the original file name when going through the loop). Then rename the file to everything after the 3rd underscore, which is the %%I value.
Your FINDSTR search is wrong - a string of any characters (wildcard) is .*, not *.
Variable find/replace does not support wildcards, except for the !var:*search=! syntax that replaces everthing up until the first occurrence of "search".
There is no need for FINDSTR, all you need is DIR with normal wildcard masking.
You can use FOR /F to parse the name into tokens. I use two loops - the first to get the entire name, and the second to parse out the portion after the 3rd _.
The following should work:
#echo off
for /f "eol=: delims=" %%A in (
'dir /b /a-d *_*_*_*'
) do for /f "tokens=3* delims=_" %%B in ("%%A") do ren "%%A" "%%C"
Or you could use my jren.bat utility that renames files using regular expression replacement. It is a hybrid JScript/batch script that runs natively on any Windows machine from XP onward.
jren "^(.*?_){3}" ""
Use CALL JREN if you put the command within another batch script.

Exclusion of files of certain extension

In a Windows cmd batch script named my.bat, I want to execute a command for all the files in the current directory except for the batch file my.bat.
I use below command in my.bat currently to run the command only for *.txt *.ppt, but really as new files of different extensions might be added to the folder and hence execution of this command by excluding one type of file extension (in my case *.bat) would be more readable/helpful.
FOR /F "delims=|" %%i IN ('dir /b *.txt *.ppt') DO echo %%i
Question is: How do I exclude that file alone with that particular file extension and make the for command execute on all files of all extensions except the excluded one?
FOR /F "tokens=* delims=|" %%i IN ('dir /b *.*') do if not %%~xi==.bat echo %%i
I have added tokens=* in as well otherwise you won't get full filenames if they have spaces.
To echo without the dot
setlocal enabledelayedexpansion
FOR /F "tokens=* delims=|" %%i IN ('dir /b *.*') do (
set e=%%~xi
set e=!e:.=!
echo !e!
)
This is providing that the file doesn't have any other dots, otherwise it will remove them too. This is a bit more sturdy than just removing the 4th character from the end though, as not all files have a 3 character extension.
You could pass the output of the dir /b command through findstr, like so:
FOR /F "delims=|" %%i IN ('dir /b ^| findstr /v /r "^my.bat$"') DO echo %%i
The /v option to findstr prints anything that doesn't match the parameter. The match is based on a regular expression that matches only lines that contain my.bat and nothing else.

Why is the FOR /f loop in this batch script evaluating a blank line?

I'm trying to write a batch script that obtains (among other things) a list of all of the disk drives the computer has. The basic code looks something like this:
REM Build the list of disk drives to monitor
SETLOCAL enabledelayedexpansion
FOR /f "skip=1 tokens=1 delims=:" %%a in ('"WMIC logicaldisk WHERE drivetype=3 GET deviceid"') do (
SET "DISK_DATABASES=!DISK_DATABASES!%%a|"
SET "DRIVES_TO_MONITOR=!DRIVES_TO_MONITOR!%%a:\\|"
)
I pretty obviously build two lists with slightly different formats for use later. When I run this, however, the output I get looks something like this:
C|D|E||
C:\\|D:\\|E:\\|:\\|
Now, I expect the trailing pipe in both cases and I can manage that, but I'm really confused why there is an extra blank entry in there. If I run the wmic command manually, I can see that there is indeed a blank line at the end of the output, but my understanding is that /f was specifically supposed to ignore blank lines.
If I turn ECHO on, it looks like that last line is just coming in as a carriage return/newline or similar. Is there a way to do what I'm expecting? Am I missing something? I tried to write an if condition in the loop to exclude this last line, but it was... funky and never worked. I appreciate any/all help.
I just came over this topic. I've been using findstr /v to exclude empty lines:
FOR /f "usebackq skip=1 tokens=1 delims=:" %%a in (`WMIC logicaldisk WHERE "drivetype=3" GET deviceid ^| findstr /v /r "^$"`) do (
In this case the last iteration produces not an empty item, and you get your output of C|D|E|| only with echo %DISK_DATABASES%,
but echo !DISK_DATABASES! will output ||D|E|??
That's because the last element is a single <CR> character.
And <CR> characters are directly removed after the percent expansion, but not with delayed expansion.
You could avoid this, using the percent expansion to remove them
setlocal EnableDelayedExpansion
FOR /f "skip=1 tokens=1 delims=:" %%a in ('"WMIC logicaldisk WHERE drivetype=3 GET deviceid"') do (
set "item=%%a"
call :removeCR
if not "!item!"=="" (
SET "DISK_DATABASES=!DISK_DATABASES!!item!|"
SET "DRIVES_TO_MONITOR=!DRIVES_TO_MONITOR!!item!:\\|"
)
)
goto :eof
:removeCR
:removeCR
set "Item=%Item%"
exit /b
According to http://ss64.com/nt/for_f.html
Many of the newer commands and utilities (e.g. WMIC) output text files in unicode format, these cannot be read by the FOR command which expects ASCII.
To convert the file format use the TYPE command.
So it appears that WMIC and FOR don't play nice together.
I discovered a more efficient and more reliable method to strip the unwanted <CR> from the end of each line. No temp file, and no CALL needed.
I don't understand the mechanism of how FOR /F converts the WMIC unicode output into ASCII. Normally FOR /F cannot read unicode. But however it works, each converted line ends with <CR><CR><LF>. FOR /F breaks lines at each <LF>, and then if the last character in the line is <CR> it strips that last <CR>, in this case leaving behind the unwanted <CR>.
The solution is to simply pass each line through one more FOR /F :-)
#echo off
setlocal enableDelayedExpansion
for /f "skip=1 delims=" %%A in (
'wmic logicaldisk where "drivetype=3" get deviceid'
) do for /f "tokens=1 delims=:" %%B in ("%%A") do (
set "disk_databases=!disk_databases!%%B|"
set "drives_to_monitor=!drives_to_monitor!%%B:\\|"
)
This method is more reliable then using normal expansion because you don't have to worry about quoting or escaping special characters. For example, The CALL method that uses normal expansion cannot handle a string like "this & that" & the other. But this method has no problem with such a string.
Add ^| findstr . and you will get only not blank lines
REM Build the list of disk drives to monitor
SETLOCAL enabledelayedexpansion
FOR /f "skip=1 tokens=1 delims=:" %%a in (
'"WMIC logicaldisk WHERE drivetype=3 GET deviceid" ^| findstr .') do (
SET "DISK_DATABASES=!DISK_DATABASES!%%a|"
SET "DRIVES_TO_MONITOR=!DRIVES_TO_MONITOR!%%a:\|"
)
My standard idiom for dealing with this is to write the output from WMIC to a temp file, then use TYPE (which reduces UTF16 to ASCII) to feed that into FOR, like this:
:: Standard environment setup
setlocal enabledelayedexpansion
:: Every variable whose name starts with "tf" will identify a temporary
:: file - remove any such variables inherited from the parent environment
for /f %%V in ('set tf') do set %%V=
:: Create some temporary filenames. Prefix all of them with this script's
:: own name to avoid clashes with those owned by other scripts.
for /l %%I in (1,1,4) set tf%%I="%temp%\%~n0-temp%%I.txt"
:: Use temp file to work around coding mismatch between WMIC out and FOR in
wmic product where "name like 'Microsoft Office %% 2010'" get packagecache >!tf1!
for /f "skip=1" %%P in ('type !tf1!') do if exist "%%~P" msiexec /x "%%~P" /passive /norestart
:: Before quitting script, clean up temporary files
for /f %%V in ('set tf') do if exist "%%~V" del /f /q "%%~V"
endlocal
Run the following command:
wmic blah /value | find "=" >> wherever
Output will be:
field=value
Note that there will be no extra lines.

How do you loop through each line in a text file using a windows batch file?

I would like to know how to loop through each line in a text file using a Windows batch file and process each line of text in succession.
I needed to process the entire line as a whole. Here is what I found to work.
for /F "tokens=*" %%A in (myfile.txt) do [process] %%A
The tokens keyword with an asterisk (*) will pull all text for the entire line. If you don't put in the asterisk it will only pull the first word on the line. I assume it has to do with spaces.
For Command on TechNet
If there are spaces in your file path, you need to use usebackq. For example.
for /F "usebackq tokens=*" %%A in ("my file.txt") do [process] %%A
From the Windows command line reference:
To parse a file, ignoring commented lines, type:
for /F "eol=; tokens=2,3* delims=," %i in (myfile.txt) do #echo %i %j %k
This command parses each line in Myfile.txt, ignoring lines that begin with a semicolon and passing the second and third token from each line to the FOR body (tokens are delimited by commas or spaces). The body of the FOR statement references %i to get the second token, %j to get the third token, and %k to get all of the remaining tokens.
If the file names that you supply contain spaces, use quotation marks around the text (for example, "File Name"). To use quotation marks, you must use usebackq. Otherwise, the quotation marks are interpreted as defining a literal string to parse.
By the way, you can find the command-line help file on most Windows systems at:
"C:\WINDOWS\Help\ntcmds.chm"
In a Batch File you MUST use %% instead of % : (Type help for)
for /F "tokens=1,2,3" %%i in (myfile.txt) do call :process %%i %%j %%k
goto thenextstep
:process
set VAR1=%1
set VAR2=%2
set VAR3=%3
COMMANDS TO PROCESS INFORMATION
goto :EOF
What this does:
The "do call :process %%i %%j %%k" at the end of the for command passes the information acquired in the for command from myfile.txt to the "process" 'subroutine'.
When you're using the for command in a batch program, you need to use double % signs for the variables.
The following lines pass those variables from the for command to the process 'sub routine' and allow you to process this information.
set VAR1=%1
set VAR2=%2
set VAR3=%3
I have some pretty advanced uses of this exact setup that I would be willing to share if further examples are needed. Add in your EOL or Delims as needed of course.
Improving the first "FOR /F.." answer:
What I had to do was to call execute every script listed in MyList.txt, so it worked for me:
for /F "tokens=*" %A in (MyList.txt) do CALL %A ARG1
--OR, if you wish to do it over the multiple line:
for /F "tokens=*" %A in (MuList.txt) do (
ECHO Processing %A....
CALL %A ARG1
)
Edit: The example given above is for executing FOR loop from command-prompt; from a batch-script, an extra % needs to be added, as shown below:
---START of MyScript.bat---
#echo off
for /F "tokens=*" %%A in ( MyList.TXT) do (
ECHO Processing %%A....
CALL %%A ARG1
)
#echo on
;---END of MyScript.bat---
#MrKraus's answer is instructive. Further, let me add that if you want to load a file located in the same directory as the batch file, prefix the file name with %~dp0. Here is an example:
cd /d %~dp0
for /F "tokens=*" %%A in (myfile.txt) do [process] %%A
NB:: If your file name or directory (e.g. myfile.txt in the above example) has a space (e.g. 'my file.txt' or 'c:\Program Files'), use:
for /F "tokens=*" %%A in ('type "my file.txt"') do [process] %%A
, with the type keyword calling the type program, which displays the contents of a text file. If you don't want to suffer the overhead of calling the type command you should change the directory to the text file's directory. Note that type is still required for file names with spaces.
I hope this helps someone!
The accepted answer is good, but has two limitations.
It drops empty lines and lines beginning with ;
To read lines of any content, you need the delayed expansion toggling technic.
#echo off
SETLOCAL DisableDelayedExpansion
FOR /F "usebackq delims=" %%a in (`"findstr /n ^^ text.txt"`) do (
set "var=%%a"
SETLOCAL EnableDelayedExpansion
set "var=!var:*:=!"
echo(!var!
ENDLOCAL
)
Findstr is used to prefix each line with the line number and a colon, so empty lines aren't empty anymore.
DelayedExpansion needs to be disabled, when accessing the %%a parameter, else exclamation marks ! and carets ^ will be lost, as they have special meanings in that mode.
But to remove the line number from the line, the delayed expansion needs to be enabled.
set "var=!var:*:=!" removes all up to the first colon (using delims=: would remove also all colons at the beginning of a line, not only the one from findstr).
The endlocal disables the delayed expansion again for the next line.
The only limitation is now the line length limit of ~8191, but there seems no way to overcome this.
Or, you may exclude the options in quotes:
FOR /F %%i IN (myfile.txt) DO ECHO %%i
Here's a bat file I wrote to execute all SQL scripts in a folder:
REM ******************************************************************
REM Runs all *.sql scripts sorted by filename in the current folder.
REM To use integrated auth change -U <user> -P <password> to -E
REM ******************************************************************
dir /B /O:n *.sql > RunSqlScripts.tmp
for /F %%A in (RunSqlScripts.tmp) do osql -S (local) -d DEFAULT_DATABASE_NAME -U USERNAME_GOES_HERE -P PASSWORD_GOES_HERE -i %%A
del RunSqlScripts.tmp
If you have an NT-family Windows (one with cmd.exe as the shell), try the FOR /F command.
The accepted anwser using cmd.exe and
for /F "tokens=*" %F in (file.txt) do whatever "%F" ...
works only for "normal" files. It fails miserably with huge files.
For big files, you may need to use Powershell and something like this:
[IO.File]::ReadLines("file.txt") | ForEach-Object { whatever "$_" }
or if you have enough memory:
foreach($line in [System.IO.File]::ReadLines("file.txt")) { whatever "$line" }
This worked for me with a 250 MB file containing over 2 million lines, where the for /F ... command got stuck after a few thousand lines.
For the differences between foreach and ForEach-Object, see Getting to Know ForEach and ForEach-Object.
(credits: Read file line by line in PowerShell )
Modded examples here to list our Rails apps on Heroku - thanks!
cmd /C "heroku list > heroku_apps.txt"
find /v "=" heroku_apps.txt | find /v ".TXT" | findstr /r /v /c:"^$" > heroku_apps_list.txt
for /F "tokens=1" %%i in (heroku_apps_list.txt) do heroku run bundle show rails --app %%i
Full code here.
To print all lines in text file from command line (with delayedExpansion):
set input="path/to/file.txt"
for /f "tokens=* delims=[" %i in ('type "%input%" ^| find /v /n ""') do (
set a=%i
set a=!a:*]=]!
echo:!a:~1!)
Works with leading whitespace, blank lines, whitespace lines.
Tested on Win 10 CMD

Resources