DosTips.com

Hello everyone.

After struggling for a little while with a batch that is printing the wrong lines I did some tests and found that for some reason 'find' and 'findstr' are giving me different line numbers for a specific line. Has anyone seen this?

c:\Temp\S>type supportshow.txt | find /N "portshow 0"
[14580]portshow 0

c:\Temp\S>type supportshow.txt | findstr /R /I /B /N /C:"portshow 0"
14320:portshow 0

if I open the file with Notepad++ the correct line number is the one given by findstr, but if then I try to print that line it prints another line entirely different. And the thing is, I found a loop somewhere on the internet that does this print specific line thing, but it works with 'find' that's why I think its printing the wrong line.

The for loop I'm using to print the lines is:
for /f "tokens=1,* delims=]" %v in ('find /n /v "" ^< "supportshow.txt" ^| findstr "^\[14580\]" ') do echo %w
(I don't fully understand this line so I haven't been able to change it, 14580 is the line I'm trying to print)

So... does anyone know why find and findstr give different line numbers? is there a modifier that I could use so that ignore whatever is making then count differently? the source file was created by having putty save the output of an ssh session (perhaps putty is adding an invisible character or something?)

or, does anyone know of another way to print a specific line of a file that doesn't rely on 'find'?

Thanks!

Hi Spodermon!

Happy Christmas and welcome to the forum.

Paste the following (untested) code into an empty Notepad window and save the file as lineNo.cmd:



Run the file and it will display the line number of the last line in "supports.txt" containing the string "ports 0". I’m assuming "supports.txt" is in the "%tmp%" folder.

Switches for findstr (and xcopy) can be grouped. Findstr also accepts a file name as argument. No need to use pipes or redirection.

HTH!

- SB

Thanks Sponge Belly.

I did as you suggested, and the output of the script find the line on 14320:

c:\Temp\S>lineno
match found on line 14320

as a test, I modified the script you shared to use 'find' instead 'findstr', and again its giving me a different result...

(after the modification:)
c:\Temp\S>lineno
match found on line [14580]portshow 0

the modification was just change findstr with find:


So, I'm still at a loss, why are they giving different outputs and how do I make them be the same. I'm guessing there's something wrong on the source file, probably some character find is interpreting as a line break and findstr isn't? any way to "clean" the file?

Maybe someone is thinking "why do you need them to be same, pick one and continue working with it!"
I need findstr because it has a modifier to look for the text at the beginning of the line, and I need find because that's the way I have to print a specific line. if anyone has an alternative to any of those things I'm good with it.

Thanks!

Ok, so I went digging into the source file and I found one of the "problem lines".

if I open the source file on notepad it looks ok, but if I open it in Notepad++ a "nul" icon is displayed. before this line both find and findstr match all the line numbers, after this line they are skewed by one line (skewed? is that the right word? sorry, not English native speaker here).

The forum does not allow me to upload a txt file, but I'm attaching an screenshot of the problematic line as seen on notepad vs Notepad++, also, adding the lines as code, just in case it helps somehow...



any ideas how to clean the file?

If you are just looking for an easy solution just remove the lines with nul using notepad++ and replace with the enter key or what you want.
By the way you can just compress the txt file in zip file, then you can upload it here.

Thanks Sounak.

I'm indeed looking for an easy solution, but I'm looking for something that can be run automatically, on the same script if at all posible

The script I'm trying to do will extract information from these kind of source files, do some processing and give a csv file for easy reading, but the source files are usually created the same way, by running a command on an ssh session and saving the output so this "nul" lines will most likely by an ongoing thing.

Any ideas?

Forgot to attach file.

Dave's JREPL utility can certainly remove the null bytes
viewtopic.php?f=3&t=6044


Steffen

Thanks aGerman!

I found another workaround.

I found that powershell counts the lines the same way findstr does, so I'm using it to print the lines I wanted to print. I added this to my script:


Where !portline! is the first line I want to print and !dife! holds the number of lines I want to print. Its also way faster. instead of doing a loop and printing line by line like i was doing, PS just throws everything in a single pass.

Hopefully using PS in a batch wont come back and bite me in the ...

Thanks everyone.

As long as you're running it on a computer that's running XP or later, you'll be fine with a batch/powershell hybrid.

Hi Again,

Null characters are notoriously difficult to deal with, as discussed in Robust Line Counter. If an input text file is contaminated with null characters, the best course of action is to get rid of them. But if that isn’t possible, there are workarounds.

For example, the following snippet accepts 2 arguments: a filename, and a case-insensitive literal string to search for.



This is a hugely inefficient way to search a file, but its one redeeming quality is that it won’t be tripped up by the presence of null characters.

Best wishes to all for a happy and peaceful 2017!

- SB