Extract text between two XML tags using sed

Question

I have XML file similar to the following:

<?xml version="1.0" encoding="UTF-8"?>
<OnlineCommentary>
    <doc docid="cnn_210085_comment002" articleURL="http://www.cnn.com/News.asp?NewsID=210085" date="10/07/2010" time="00:21" subtitle="Is Justin Bieber getting special treatment?" author="Zorro75">
        <seg id="1"> They are the same thing. Let's shoot them both. </seg>
    </doc>
    <doc docid="cnn_210092_comment004" articleURL="http://www.cnn.com/News.asp?NewsID=210092" date="06/04/2010" time="17:07" subtitle="Dear Chicago, we love you despite it all" author="MRL1313">
        <seg id="1"> We can't wait for you to move back either. </seg>
        <seg id="2"> You seem quite uptight. </seg>
        <seg id="3"> Does your wife (who is also your sister) not give it up any more? </seg>
    </doc>
</OnlineCommentary>

I would like to to execute command on this file to extract only the contnet between the opening tag <seg ...> and the closing tag </seg>

I tried :

sed -n 's:.*<seg id="1">\(.*\)</seg>.*:\1:p' XML-file.xml > output.txt

My questions are the following:

-- How can I print all <seg id="*"> ?? my command prints only the the content of the first tag ( <seg id="*"> )

-- Is that is there a way that can be used to make for example the <seg id="1"> , <seg id="2"> , <seg id="3"> to be printed in the same line while the tag that include only <seg id="1"> to be printed in separate line??

Answer 1

Use a proper XML handling tool. For example, in XML::XSH2 :

open file.xml ;
for //doc echo seg/text() ;

Answer 2

print all the <seg id=> (one per line) including <seg

sed -n 's:.*\(<seg id="[0-9]\{1,\}">.*</seg>\).*:\1:p' XML-file.xml > output.txt

Print all on 1 line with separated , . Use of holding buffer instead of printing and at the end, recall the buffer, replace new line by , (and remove starting , due to Append action), and print the result

sed -n '\:.*\(<seg id="[0-9]\{1,\}">.*</seg>\).*:  { s//\1/
   H
   }
$ {g
   s/\n/,/g;s/^,//
   p
   }' XML-file.xml > output.txt

Now, the advice of @Choroba to use adequat XML tools is very good, you minimize the risk of treating unwanted data of the file.