Adding new tag to an xml file using sed

nano2 · 11-17-2011, 10:25 AM

Hi

Have an xml file where by i want to add a tag and its value to it

BEFORE :

Code:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE application PUBLIC
        "-//Sun Microsystems, Inc.//DTD"
        "http://java.sun.com/dtd/ab.dtd">


<app>

    <bbb>
        <jjj>test1</jjj>

     </bbb>

     <bbb>   
       <jjj>test2</jjj>
    </bbb>

    <bbb>
        <jjj>test3</jjj>
    </bbb>

</app>

AFTER :

Code:

</app><?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE application PUBLIC
        "-//Sun Microsystems, Inc.//DTD"
        "http://java.sun.com/dtd/ab.dtd">

<app>

    <bbb>
        <jjj>test1</jjj>

     </bbb>

     <bbb>   
       <data>test2</data>
    </bbb>


    <bbb>   
       <jjj>test2</jjj>
    </bbb>
       
   

    <bbb>
        <jjj>test3</jjj>
    </bbb>
  

</app>

i was trying to do was to use sed to add the new tag <data>

Code:

FILE=~/test.xml

sed -i.bck 's#\(app\).*\(;/app\)#\1<bbb>   
       <data>test2</data>
    </bbb>\2#g' $FILE

This however does not work any ideas ?

David the H. · 11-17-2011, 11:03 AM

Could you explain in more detail your exact requirements? why are you trying to substitute "app", etc...? Are you really trying to run the substitution on the whole section?

sed is not a very good multi-line editor, and only operates on one line at a time by default. Targeting multi-line blocks often requires complex handling.

What we really need to know is, what defines the line that needs to be changed?

If it's the content of the tag, for example, you may really only need something like this:

Code:

sed '\|test2| s|jjj|data|g' file

Or if you need it to be more accurate, and specifically target the tag itself, perhaps this:

Code:

sed -r '\|test2| s|<(/?)[^>]+>|<\1data>|g' file

But if you need it to match something only within a certain sub-section of lines or something, that's more difficult, and maybe even close to impossible. So again, we need you do define the problem more exactly. If the necessary operation is much more complex than a simple text-matching, then you may have to turn to a different tool that's capable of doing true xml parsing and editing.

Considering that xml structure is free-form and can easily span multiple lines, that may be a good idea in any case.

nano2 · 11-18-2011, 03:14 AM

I want to insert the tag below there is no replacement needed I want the tag below slotted in betweenthe tags <app> </app>

<bbb>
<data>test2</data>
</bbb>

David the H. · 11-18-2011, 12:25 PM

Ah, ok. I failed to see that it was a separate insert. I thought you were just modifying the tags already there.

sed has the dedicated commands -i (insert before), and a (append after) for adding lines of text to a file, instead of substitution. The following will append the desired section directly after the <app> tag.

Code:

sed '\|<app>| a \
\n\t<bbb>\n\t\t<data>test2</data>\n\t</bbb>
' file

If you need it to go in exactly the place shown in the output above, then you need to use nesting to target the first </bbb> after <app>.

Code:

sed '\|<app>|,\|</app>| { \|</bbb>| a \
\n\t<bbb>\n\t\t<data>test2</data>\n\t</bbb>
}' file

Note the multi-line nature of the a and i text. You need to separate out the text block to insert with the initial slash+newline and final raw newline (also backslash any other literal newlines before the last one). gnu sed is a little more flexible than some other versions on this, but for larger blocks of text it's probably safer just to use them anyway.

http://www.grymoire.com/Unix/Sed.html
http://sed.sourceforge.net/grabbag/
http://sed.sourceforge.net/sedfaq.html
http://sed.sourceforge.net/sed1line.txt

I still say it would be better to use a tool that's designed specifically for editing xml. This one looks like it may be useful, for example:

http://xmlstar.sourceforge.net/overview.php

Or at the very least you could the file through xmllint to clean up the formatting first.

nano2 · 11-22-2011, 08:38 AM

Code:

sed '\|<app>|,\|</app>| { \|</bbb>| a \
\n\t<bbb>\n\t\t<data>test2</data>\n\t</bbb>
}' file

This appended after each <bbb> tag element i just want it enter once ....

David the H. · 11-23-2011, 05:35 AM

Well then, this just illustrates exactly why sed isn't the proper tool to use. It's not really designed for editing things like xml files. How many times do I have to suggest that you find a true xml parser for this?

I don't know why I was thinking it would only append to the first instance, but I certainly was wrong about it. I can't think now, offhand, of any easy way to limit it to only one match. Sorry.

jschiwal · 11-23-2011, 05:47 AM

Code:

sed '/<bbb>/,/<\/bbb>/{
                        /<jjj>test2<\/jjj>/s#<jjj>test2</jjj>#<data>test2</data>#
                      }' test.xml

To select a range of lines, use pattern selection: /pattern1/,/pattern2/{ .. }
The sed commands inside the braces will only be run inside the range of lines matching the pattern. You can have ranges
inside ranges.

You may need to add another line for the case when the xml tag is contained in the same line.

When working with xml, an xslt transform may be a better way to go.