gsub: replacing a.*a if no occurence of b in .*

The _question_ assumed that, which is why the answers did too.

Oh yes, I totally agree, the file snippet the OP provided did indeed  
assume that, though nothing in the text of his question did, so I  
wasn't entirely clear whether the actual file that is going to be  
processed has this form or not. So I just wanted to make sure the OP  
is aware of this limitation, in case the actual file is more  
problematic.

But most importantly, I wanted to suggest a reevaluation, if  
possible, of the process that generates these XML's, and perhaps  
fixing that, instead of patching the problem after it has been created.

Haris Skiadas
Department of Mathematics and Computer Science
Hanover College