Shell Programming and Scripting

I have an xml file, from where I need to take out Application2 entries and keep the others. I need to remove from <product> to </product> and the key element to look for while removing should be <application> as other pairs can be same for others.

<product>
<name>Name1</name>
<version>1.0</version>
<Location>c://location1</Location>
<application>Application1</application>
</product>
<product>
<name>Name2</name>
<version>1.0</version>
<Location>c://location2</Location>
<application>Application2</application>
</product>
<product>
<name>Name2</name>
<version>1.0</version>
<Location>/c://location2</Location>
<application>Application3</application>
</product>

Thanks,
Chiru

If the start line is unique, I can remove the pattern using:
sed -e "/<product>/,/<\/product>/d" filename > tmpfile

If I use the above, it removes all the patterns.
I am not able to figure out how to remove if the unique field is somewhere in the middle - like <application>

Please help.

Thanks,
Chiru

Not tested but anyways try this:

Code:

sed -n '
/product\>/ b block
H
$ b block
b
:block
x
/Application2/!p' filename

Thanks ,
I tried what you sent, from the result I see that it removes from <product> to </Application2>
The last line of the pattern </product> is not removed.

Thanks,
Chiru

Below code is working fine:

Code:

sed -n '
/<product\>/ b block
H
$ b block
b
:block
x
/Application2/!p'

You can give this a try:

Code:

awk 'BEGIN {ORS=RS="</product>"}{if(substr($5,14,13)!="Application2<")print}' file

Regards

Your code is not working