[SOLVED] find opposite with grep

xeon123 · 03-10-2011, 10:48 AM

Hi,

I want to print all lines that doesn't match a name in bash.
I'm using grep, but my problem is that I don't know to find the opposite of I what I want.

For example, in the text:

Code:

This
is
a
string

I want to print all line that doesn't match the word 'is'.
I want to print

Code:

This
a
string

How can I do that?

j1alu · 03-10-2011, 10:58 AM

grep -v
so this might (!) work for you:
grep -v "\bis\b"
I am unsure about the quotes. the both \b make it being a standalone word
(else you would hit "This" too)

Don't take my word for it.

siranjeevi · 03-10-2011, 11:01 AM

Hi,

The -v option will exclude the matchs,

cat filename | grep -v is

xeon123 · 03-10-2011, 11:18 AM

I'm not asking correctly.
This is my problem:

I've a tmp.txt file with this content:

Code:

My
name
is/was
test

And I'm using the following script:

Code:

cat tmp.txt | while read line
do
            if [[ $line != "^is" ]]
            then
                echo $line                                                  
            fi;
     
done

but the problem is that if clause is not excluding the "is/was" word. What's wrong?

xeon123 · 03-10-2011, 11:29 AM

I've found the solution:

Code:

cat tmp.txt | while read line
do
            if [[ $line != is* ]] // removing the ""
            then
                echo $line                                                  
            fi;
     
done

druuna · 03-10-2011, 11:31 AM

Hi,

The example posted in post #4 works on my side:

Code:

$ cat infile
My
name
is
test

$ cat foo.sh
 #!/bin/bash

cat infile | while read line
do
  if [[ $line != "is" ]]
  then
    echo $line                                                  
  fi
done

$ ./foo.sh
My
name
test

Are you sure your infile is a linux file?
file infile should give you infile: ASCII text and not infile: ASCII text, with CRLF line terminators. Use dos2unix infile to reformat from dos/windows to unix.

Hope this helps.

Original post was edited while replying....

Nylex · 03-10-2011, 11:33 AM

Another way to do it is with sed:

sed '/^is/ d' tmp.txt

will delete lines beginning with "is". If you want to change the file in place, you can use sed's -i option. A tutorial for sed can be found here.

grail · 03-10-2011, 07:18 PM

Hey druuna ... You missed the updated input file (is now is/was) and the carat (^) in front of 'is' in the script.

OP - I agree with Nylex that sed or grep would be much simpler, unless of course you are requiring to do more in your script

druuna · 03-11-2011, 01:40 AM

Quote:

Originally Posted by grail

Hey druuna ... You missed the updated input file (is now is/was) and the carat (^) in front of 'is' in the script.

Hey grail.... I noticed, hence the remark at the bottom of my post.