I have the file below:
some lines = \
some params
SUBDIRS = \
text1 \
text2 \
# commented text
text3 \
text4 \
OTHERS = \
other text here \
I want to extract a list containing text1, text2, text3, text4.
How can I proceed?
This extracts all the lines after SUBDIRS but before OTHERS excluding lines that start with # or are empty:
$ awk '/OTHERS/{f=0} f && /./ && !/^#/{print} /SUBDIRS/{f=1}' file
text1 \
text2 \
text3 \
text4 \
This program uses one variable, f: f is 1 when we are in the desired range of lines and 0 elsewhere.
/OTHERS/{f=0}
If we have reached OTHERS, then set f back to zero.
f && /./ && !/^#/{print}
If f is nonzero (f) and the line is not blank (/./) and the line does not start with # (!/^#/), then print this line.
/SUBDIRS/{f=1}
If we have reached the line containing SUBDIRS, then set f to 1.
Because print is the default action when no action is specified, we can omit it from the script:
$ awk '/OTHERS/{f=0} f && /./ && !/^#/; /SUBDIRS/{f=1}' file
text1 \
text2 \
text3 \
text4 \
This removes the trailing \ and combines all the output to one line:
$ awk '/OTHERS/{f=0} $NF=="\\"{$NF=""} f && /./ && !/^#/{a=a" "$0} /SUBDIRS/{f=1} END{print a}' file
text1 text2 text3 text4
Try this. SPACE_TAB is, I hope obviously, a space plus a tab.
$ SPACE_TAB=" "; sed -ne '/^SUBDIRS/,/^['"$SPACE_TAB"']*$/p' test.in
SUBDIRS = \
text1 \
text2 \
# commented text
text3 \
text4 \
bash, you may want to use SPACE_TAB=$'\t ' in order to be sure that the correct characters are assigned to the variable.
sed -ne '/^SUBDIRS/,/^[[:blank:]]*$/p'
grep -ow 'text[0-9]'could do the job for the example (assuming GNUgrep). If you want the file list associated withSUBDIRS, then you have to work harder — this is what the answers so far have assumed. What should the output format be? One word per line without backslashes? Does the code need to handle all the niceties ofmakesyntax? How will you invoke the command? What should happen if the name specified is not found in the file?