shellscripting to edit xml file

Shell Programming and Scripting
1

I have an xml file where every line which has the word CDATA in it follows this pattern

(line number)  <word1><![CDATA[something]]></>

I need only these lines editing so that the end result is that

(line number)  <word1><![CDATA[something]]></word1>

so it copies the first bit to the end.

Anyone know how I can do this I am having trouble with awk and sed

p.s this is in UNIX

2

You want to fetch "something" from "somethingelse" ?? :wall:
or you just want to have </word1> as the last tag?

--ahamed

3 
kent$  echo '(line number)  <word1><![CDATA[somethingelse]]></>'|sed -r '/word1/{s#(<\!\[CDATA\[).*(\]\]>)#\1something\2#;s#</>#<word1/>#}'
(line number)  <word1><![CDATA[something]]><word1/>
4

just word1 as the last tag sorry my bad it was a type

I have edited now to make amends

5

legolad,

(line number)  <word1><![CDATA[something]]></word1>

Just for making clear..

You want :

(line number)  <word1><![CDATA[something]]><word1/>
(line number)  <word2><![CDATA[something]]><word2/>
(line number)  <word3><![CDATA[something]]><word3/>

OR

(line number)  <word1><![CDATA[something]]><word1/>
(line number)  <word1><![CDATA[something]]><word1/>
(line number)  <word1><![CDATA[something]]><word1/>

which one..???

6

Sorry it is this one

(line number)  <word1><![CDATA[something]]><word1/> 
(line number)  <word2><![CDATA[something]]><word2/>
(line number)  <word3><![CDATA[something]]><word3/>

however there are other lines in the code that I don't want editing so only the ones which have CDATA in them if you understand?

also it should be

(line number)  <word1><![CDATA[something]]></word1> 
(line number)  <word2><![CDATA[something]]></word2> 
(line number)  <word3><![CDATA[something]]></word3>

not

(line number)  <word1><![CDATA[something]]><word1/> 
(line number)  <word2><![CDATA[something]]><word2/> 
(line number)  <word3><![CDATA[something]]><word3/>
7

One more doubt, is the tag name "word" constant? or will it keep changing to some other tag name?

And this is what you want?

Input

line number)  <word1><![CDATA[something]]></>     
(line number)  <word1><![CDATA[something]]></>
(line number)  <word1><![CDATA[something]]></>
(line number)  <word1><![CDATA[something]]></>

Output

(line number)  <word1><![CDATA[something]]></word1>     
(line number)  <word2><![CDATA[something]]></word2>
(line number)  <word3><![CDATA[something]]></word3>
(line number)  <word4><![CDATA[something]]></word4>

--ahamed

8

No the tag name word changes from line to line

the code at the moment is

(line number)  <word1><![CDATA[something]]></> 
(line number)  <word2><![CDATA[something]]></> 
(line number)  <word3><![CDATA[something]]></> 
(line number)  <word4><![CDATA[something]]></>

But it needs to end up as

(line number)  <word1><![CDATA[something]]></word1>      
(line number)  <word2><![CDATA[something]]></word2> 
(line number)  <word3><![CDATA[something]]></word3> 
(line number)  <word4><![CDATA[something]]></word4>

also the something tag word changes each line as well

9

Try this

awk -F"<" '{t=length($2);x=substr($2,0,t-1);sub(/\/>/,"/"x">");print}' input_file

--ahamed

---------- Post updated at 07:58 AM ---------- Previous update was at 07:57 AM ----------

If solaris, please use nawk!

--ahamed

11

worked a dream, thanks a lot for your time

12

great work ahamed..

test$ cat new7 
(line number)  <word1><![CDATA[something]]></>
(line number)  <word2><![CDATA[something]]></>
(line number)  <word2><![CDATA[something]]></>


test$ cat new7 |awk -F"<" '{t=length($2);x=substr($2,0,t-1);sub(/\/>/,"/"x">");print}'
(line number)  <word1><![CDATA[something]]></word1>
(line number)  <word2><![CDATA[something]]></word2>
(line number)  <word2><![CDATA[something]]></word2>