File content checking and replacing it!

[Pando]

I have a file which contain the line
...
...
"-"
...
...
I need to get the "-" part to be passed to a variable ad subsequently replaced it by "9999". How can i get it in a single command?

Maximum points for correct reply!

[H.Merijn Brand (procura · Accepted solution]

A variable to the changing process, or to the calling shell?

# perl -pi -e's{"(.*?)"}{"9999"} and $perl_var = $1' file

would change the content of file and have the variable available just after the change in perl as $perl_var

If it is only one line that is changed, and you want to promote the variable to the shell, you need to catch it and set it

# MY_VAR=`(perl -pi -e's{<(SUBID)>"(.*?)"}{<$1>"9999"} and print STDERR "$2\n"' file) 2>&1`

As a proof of concept:
--8<---
sh-2.05b$ MY_VAR=2
sh-2.05b$ echo $MY_VAR
2
sh-2.05b$ cat file
Line 1

Line 3
"1234"

Line 6
sh-2.05b$ MY_VAR=`(perl -pi -e's{<(SUBID)>"(.*?)"}{<$1>"9999"} and print STDERR "$2\n"' file) 2>&1`
sh-2.05b$ echo $MY_VAR
1234
sh-2.05b$ cat file
Line 1

Line 3
"9999"

Line 6
sh-2.05b$
-->8---

Enjoy, Have FUN! H.Merijn

[Elmar P. Kolkman]

For the replacing you simply can use sed:
sed 's|".*"|"9999"'

But since you also want the old value in a variable, it gets trickier.

You could do something like this:

oldvalues=$(sed -n '/".*"<\/SUBID>/s|.*"\(.*\)".*|\1|p' input)
sed 's|".*"|"9999"|' input > output

This means, of course, that you run twice through the input file with sed...

[Pando]

Thanks for that quick reply.