[SOLVED] Using sed to change strings

[juliojj14]

I am having a lot of problems trying to change one string by another using sed:
the sentence is like this:
sed -i 's/KERNEL=="tty[A-Z]*", NAME="%k", GROUP="uucp", MODE="0660"/KERNEL=="tty[A-Z]*", NAME="%k", GROUP="uucp", MODE="0666"/g' 50-udev.rules

it is just to fing the line with:
KERNEL=="tty[A-Z]*", NAME="%k", GROUP="uucp", MODE="0660"
and replace it with:
KERNEL=="tty[A-Z]*", NAME="%k", GROUP="uucp", MODE="0666"
I would like to identify the whole line because other lines have the same structure.
I beliebe that between KERNEL=="tty[A-Z]*", and NAME="%k" there are 12 spaces (it is the file 50-udev.rules for red hat)
I will appreciate your help!

[nmset]

sed -i 's/KERNEL=="tty\[A-Z\]\*", NAME="%k", GROUP="uucp", MODE="0660"/KERNEL=="tty\[A-Z\]\*", NAME="%k", GROUP="uucp", MODE="0666"/g' 50-udev.rules

You just have to prevent the character substitution [A-Z] and the * from being interpreted by sed, and impose them as string litterals.

[juliojj14]

Thank you very much for your asnwer. Unfortunately, I still could not achieve to make it work. Could you help me with a working example, please?
Thank you very much again!

[nmset]

... It seems to be quite straight forward ! You should perhaps feed in the full path to 50-udev.rules. Are you on Linux ? or some other 'NIX ? Not all commands are portable from one 'NIX to the other.

[juliojj14]

I am in Linux (Red Hat). I would like to create a script instead of writting it manualy. I do not achieve to impose [A-Z] and * as string literals.

[Franklin52]

Try this:

awk -v var='KERNEL=="tty[A-Z]*", NAME="%k", GROUP="uucp", MODE="0660"' '
$0 == var{sub("MODE=\"0660\"","MODE=\"0666\"")}1' file > newfile
mv newfile file

[juliojj14]

Thank you for yor quick answer.
with this awk example, it did not work. I am trying with smaller strings to check it but it does not make it neither. Could you help me to find out how to solve the problem?
Thanks!

[Franklin52]

Works fine for me, this is the output I get with this example, note that only the first line matches the pattern:

$ cat file
KERNEL=="tty[A-Z]*", NAME="%k", GROUP="uucp", MODE="0660"
KERNEL=="tty[A-Z]*", NAME="%s", GROUP="uucp", MODE="0660"
KERNEL=="tty[A-Y]*", NAME="%k", GROUP="uucp", MODE="0660"
KERNEL=="tty[A-Z]*", NAME="%n", GROUP="uucp", MODE="0660"
$
$ awk -v var='KERNEL=="tty[A-Z]*", NAME="%k", GROUP="uucp", MODE="0660"' '
$0 == var{sub("MODE=\"0660\"","MODE=\"0666\"")}1' file
KERNEL=="tty[A-Z]*", NAME="%k", GROUP="uucp", MODE="0666"
KERNEL=="tty[A-Z]*", NAME="%s", GROUP="uucp", MODE="0660"
KERNEL=="tty[A-Y]*", NAME="%k", GROUP="uucp", MODE="0660"
KERNEL=="tty[A-Z]*", NAME="%n", GROUP="uucp", MODE="0660"
$

[juliojj14]

Ok! I used the two line commands as an onli one line:
awk -v var='KERNEL=="tty[A-Z]*", NAME="%k", GROUP="uucp", MODE="0660"' '$0 == var{sub("MODE=\"0660\"","MODE=\"0666\"")}1' file

instead of
awk -v var='KERNEL=="tty[A-Z]*", NAME="%k", GROUP="uucp", MODE="0660"' '
$0 == var{sub("MODE=\"0660\"","MODE=\"0666\"")}1' file

By the way, I just got it as well using:
sed -i 's/KERNEL=="tty\[A-Z\]\*"\,[ ][ ]*NAME="%k", GROUP="uucp", MODE="0660"/KERNEL=="tty\[A-Z\]\*"\,\t\tNAME="%k", GROUP="uucp", MODE="0666"/g' 50-udev.rules
Thank you very much for your help!!! It was really useful