Re: A Simple Text-Region Deletion Question . . .

Greetings Marcus,


On Fri, Oct 03, 2014 at 08:44:36PM -0400, 'Marcus L. Thompson' [email protected] wrote:
> Greetings to all.
> 
> I have come across what seems to be a quite simple problem involving an
> application of sed in the context of text-block deletion. Here's the
> hypothetical text block in question:
> 
> <property name="name">
> ...
> </property>
> </property>
> 
> Now, deleting all from <property name="name"> to the _/first/_
> </property> tag is understandable enough:
> 
> Code:
> 
> sed -i '/<property name="name">*$/,/<\/property>*$/d' /file.xml
> 
> However, incorporating the 2nd </property> tag in the deletion sweep is
> a bit tougher; and just that much outside my new skillset range ;o)
> 
> Any ideas as to how one might amend this "one-liner" to make it all work?
You might try:


    $ sed '/<property name="name">/{:loop;/<\/property>.*<\/property>/{s/.*//;b;};N;b loop;};' file.xml


But this is not a good approach to edit XML files. Have a look in XML
parsers like xmllint or xsh (command-line applications).