Shell Programming and Scripting

Hi hackers,

I would like to parse the value of one XML tag (e.g. <TAG>).
Following is the script that I have written.

awk -F"[<>]" '/<TAG>/,/<\/TAG>/ {
if ( NF == 1 ) {
sub(/^[ \t]*/,"", $1)
print $1
}
}

Above ascipt works fine for the following XML tag syntax. i.e. prints "Hi"

<TAG>
Hi
</TAG>

But the script is not able to handle the following cases

1. <TAG>Hi</TAG>

2. <TAG>Hi
</TAG>

3. <TAG>
Hi</TAG>

Could someone help me out to enhance the script so that script can handle the above 3 cases also.

TIA,
Viki

Code:

awk -F"[<>]" '/<TAG>/,/<\/TAG>/ { gsub("<[^>]+>","")} ; $0 ' filename

Viki,
See if this works for you:

Code:

sed -e 's/<...>//' -e 's/<....>//' -e '/^$/d' input_file

Thanks for quick response.

The suggested script is able to parse the XML tag. But suggested script prints all the lines (except XML tags) present in the file.
My requirement is to extract & print only the value of XML tag from the file.

For example,
If the file contains the following

<FILE>
www.unix.com
Enhanced Mode
<TAG> Hi
</TAG>
Enhanced Mode

</FILE>

I need to print only "Hi".


TIA,
Viki

Viki,
All your examples you only specified XML tags and its values.
We can only go by the samples you give to us.
Now you are having an entire new specification.
Give us time to think.

Hi Shell_Life,

My apologize for not giving the full config in the first mail.

Thanks,
Viki

Viki,
Not a problem -- we get all kinds of incomplete specs here.
See if this works for you:

Code:

paste -s - - < input_file | tr '<' '\n' | egrep '^TAG' | sed 's/TAG>//'

or

Code:

paste -s - - < input_file | tr '<' '\n' | sed -n 's/^TAG>\(.*\)/\1/p'