93

I have a generated txt file. This file has certain lines that are superfluous, and need to be removed. Each line that requires removal has one of two string in the line; "ERROR" or "REFERENCE". These tokens may appear anywhere in the line. I would like to delete these lines, while retaining all other lines.

So, if the txt file looks like this:

 Good Line of data bad line of C:\Directory\ERROR\myFile.dll Another good line of data bad line: REFERENCE Good line

I would like the file to end up like this:

 Good Line of data Another good line of data Good line

TIA.

Improve this question
0

5 Answers 5

Reset to default
129

Use the following:

type file.txt | findstr /v ERROR | findstr /v REFERENCE

This has the advantage of using standard tools in the Windows OS, rather than having to find and install sed/awk/perl and such.

See the following transcript for it in operation:

 C:\>type file.txt Good Line of data bad line of C:\Directory\ERROR\myFile.dll Another good line of data bad line: REFERENCE Good line C:\>type file.txt | findstr /v ERROR | findstr /v REFERENCE Good Line of data Another good line of data Good line
Improve this answer
Sign up to request clarification or add additional context in comments.

20 Comments

To be complete: you need to add "> output.txt" (without quotes) to have the output written in a file.
@StarShip3000, I'd be very surprised if piping a text file through findstr would modify the lines in it somehow. It should just remove whole lines. You might want to open a question if that's the case, I'll keep an eye open.
@paxdiablo I just tried to reproduce with a manually created test case and couldn't get it to reproduce. I know this happened, but maybe this command against my data wasn't the cause. I'll consider myself wrong until I can prove otherwise.
thanks, this helped me out but a word of caution, I had some issues with phrases including spaces and had to use /c: so the command would look like this: type file.txt | findstr /v /c:"ERROR" | findstr /v /c:"REFERENCE". Also if you need to strip lines off that you know are at the top fo the file, pipe it through more +<n> where <n> is the number of lines to strip off.
why is it deleting everything in my file test.java if I write type test.java | findstr /v "ECHO is on." > test.java but it do what is expected if I write it to a new file like type test.java | findstr /v "ECHO is on." > test2.java
|
63

You can accomplish the same solution as @paxdiablo's using just findstr by itself. There's no need to pipe multiple commands together:

findstr /V "ERROR REFERENCE" infile.txt > outfile.txt

Details of how this works:

  • /v finds lines that don't match the search string (same switch @paxdiablo uses)
  • if the search string is in quotes, it performs an OR search, using each word (separator is a space)
  • findstr can take an input file, you don't need to feed it the text using the "type" command
  • "> outfile.txt" will send the results to the file outfile.txt instead printing them to your console. (Note that it will overwrite the file if it exists. Use ">> outfile.txt" instead if you want to append.)
  • You might also consider adding the /i switch to do a case-insensitive match.
Improve this answer

6 Comments

type !INSTALLED_APPS! | findstr /v "@ //" > temp_filter_out_junk.txt >nul 2>&1
I tried this type D:/e5/datastream/package.json | findstr /v eikon-framework | findstr /v df-core but it doesn't delete the lines
I tried findstr /V "eikon-framework df-core" D:\e5\datastream\package.json > D:\e5\datastream\package.json but it deletes all the contents if I try replace the data in the same file and If I use different out put file name then it works. How do I update the same file?
@kittu Use ">>" instead of ">". From the answer: Use ">> outfile.txt" instead if you want to append. You can't alter the same file in that single command.
Well since the question is about replacing in a file, how do you alter the current file? Everything I tried just leads to the file being empty
|
7

If you have sed:

sed -e '/REFERENCE/d' -e '/ERROR/d' [FILENAME]

Where FILENAME is the name of the text file with the good & bad lines

Improve this answer

1 Comment

Don't you need the -i flag?
1

If you have perl installed, then perl -i -n -e"print unless m{(ERROR|REFERENCE)}" should do the trick.

Improve this answer

Comments

0

It seems that using the FIND instead of the FINDSTR can support also unicode characters. e.g. type file.txt | find /v "Ω"

Improve this answer

Comments

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.