Question

I need to parse a file that has the following formatted data and get the
DIRNAME for the OTHER node.

    CLASS=
     (SOURCE=
        (TYPE=FILE)
        (DEFAULT=
           (DIRNAME=${HOME}/information/logs)
        )
      )


    OTHER=
     (SOURCE=
        (TYPE=FILE)
        (DEFAULT=
           (DIRNAME=${HOME}/site/location)
        )
      )

    STUDENT=
     (SOURCE=
        (TYPE=FILE)
        (DEFAULT=
           (DIRNAME=/opt/students)
        )
      )

I have to capture everything that is contained in the OTHER= field, like
so:

    OTHER= <whitespace> ( <to capture> ) 

and then I have to capture everything that is within the DIRNAME in the
OTHER= field, like so:

    (DIRNAME=<to capture>)

I would like to do this in a sort of robust script that can run in most
unix systems, anyone know what command line tools I should use for this
and what sort of regular expressions would I need to capture the data in
the way I have outlined.

Any help appreciated,

Ted

Answer 1

see the test below:

kent$  cat t
    CLASS=
     (SOURCE=
        (TYPE=FILE)
        (DEFAULT=
           (DIRNAME=${HOME}/information/logs)
        )
      )


    OTHER=
     (SOURCE=
        (TYPE=FILE)
        (DEFAULT=
           (DIRNAME=${HOME}/site/location)
        )
      )

    STUDENT=
     (SOURCE=
        (TYPE=FILE)
        (DEFAULT=
           (DIRNAME=/opt/students)
        )
      )

kent$  awk -F= '$1~/OTHER/{i++;print $2} $1~/DIRNAME/ && i{i=0;gsub(/\)$/,"",$2); print $2}' t

${HOME}/site/location

note that there was an empty line above output line ${HOME}/site/location,
it was the whitespace after OTHER=