Shell Programming and Scripting

Hi,
Could someone tell me how to delete a particular text inside an existing file without opening the file?
e.g. I have a text file called mytext.txt which contains three lines of text below and I want to delete “ ;” and delete the second line “b ;” including the carriage return.

a ;
b ;
c ;

So I want the file to have the contents below:

a
c

Any help will be appreciated.
p.s. I can do this in vi by running “%s/ ;//g” and deleting the second line with dd but I want this done without opening the file.

Try:

Code:

sed -e '/b ;/d' mytext.txt > mytext.txt.new

Regards,
Nir

Code:

sed -e "s/;$//g" -e "/b/d" mytext.txt > mytext.txt.new

Vino

Thanks Nir!

That command did delete the second line but it still contained " ;" for lines one and three.

I added an extra command to get rid of the " ;" for lines one and two as well.

sed -e '/b ;/d' mytext.txt | sed -e 's/ ;//' > mytext.txt.new


Thanks for your help.

Cheers
Steve

Avoid using sed twice. Instead give two different scripts with one sed.

Give it as

Code:

sed -e "s/;$//g" -e "/b/d" mytext.txt > mytext.txt.new

vino

Thanks Vino,

I have another question.

Could someone tell me how to delete a string from a line in a specific position using sed or other editors which doesn't require files to be opened?

e.g. I have a file called myfile.txt which contains three lines of text below and I want all of the lines to have 5 characters or less.

1234567
a b c d e
098


I want the result to be like below:

12345
a b c
098

Any help will be appreciated.

BTW please let me know if you have any recommended sites for learning unix commands such as above.

Cheers
Steve

Quite a number of ways to do it.

Here are two ways using sed

Code:

sh-2.05b$ echo "1234567
> a b c d e 
> 098" | sed -e "s/\(.....\).*/\1/g"
12345
a b c
098



sh-2.05b$ echo "1234567
> a b c d e 
> 098" | sed -e "s/\(.\{5\}\).*/\1/g"
12345
a b c
098

These are not the only 2 ways using sed. There could be other ways as well.

vino