Get substring from file using "sed"

Question

Can anyone help me to get substring using sed program?

I have a file with this line:

....
define("BASE", "empty"); # there can be any string (not only "empty").
....

And I need to get "empty" as string variable to my bash script.

At this moment I have:

sed -n '/define(\"BASE\"/p' path/to/file.ext
# returns this line:
# define("BASE", "empty");
# but I need 'empty'

UPD: Thanks to @Jaypal

For now I have bash script:

DBNAME=`sed -n '/define(\"BASE\"/p' path/to/file.ext`
echo $DBNAME | sed -r 's/.*"([a-zA-Z]+)".*/\1/'

It work OK, but if there any way to make the same manipulation with one line of code?

Answer 1

You should use is

sed -n 's/.*\".*\", \"\(.*\)\".*/\1/p' yourFile.txt

which means something (.*) followed by something in quotes (\".*\"), then a comma and a blank space (,), and then again something within quotes (\"\(.*\)\").

The brackets define the part that you later can reuse, i.e. the string within the second quotes. used it with \1.

I put -n front in order to answer the updated question, to get online the line that was manipulated.

Answer 2

This should help -

sed -r 's/.*"([a-zA-Z]+)"\);/\1/' path/to/file.ext

If you are ok with using awk then you can try the following -

awk -F\" '/define\(/{print $(NF-1)}' path/to/file.ext

Update:

DBNAME=$(sed -r '/define\(\"BASE\"/s/.*"([a-zA-Z]+)"\);/\1/' path/to/file.ext)

Answer 3

 sed -nr '/^define.*"(.*)".*$/{s//\1/;p}' path/to/file.ext