Hi,

I have been a long time lurkerer and admire what many of you do.

I have a little conundrum,
I would like to be able to change this text (all in one line):
<campaign><cluster><S_NAME>S-cpu</S_NAME><list><A_NODE name=P-3 ltances=1/><A_NODE name=P-4 ltances=2/></list></cluster></campaign>

into the following which has double quotes next for the values defined for 'name' and 'ltances':
<campaign><cluster><S_NAME>S-cpu</S_NAME><list><A_NODE name="P-3" ltances="1"/><A_NODE name="P-4" ltances="2"/></list></cluster></campaign>

What makes this harder is that I have no way of predicting the names used (though it will be quite similar) used for ltance and name, and there can more the just P-3 or P-4 (such as P-5, P-6, etc.)

My knowledge of sed is very limited (simple substitutions) and I don't know if its possible to do so, especially considering the whole text is in one line.

Hi and welcome to LinuxQuestions!

Try this:
and feel free to ask for clarifications.

Thanks a bunch!
That did the trick

A slight variation:

1 members found this post helpful.

A small correction to grail's post. Since the expression includes a forward slash, you have to change the delimiter to something else.

Any ascii character will do. The first character after the 's' will be used as the delimiter for the whole expression. I used '%' here.


(Edit: Hmmm... has sed gotten smarter? It appears not to complain as long as the character in question is inside '[]' brackets. At least the gnu version doesn't. I don't remember it working like that before.)


BTW, the above simply adds quotes around anything that appears between an equals-sign and either a space or slash. To avoid false positives you may want to further limit the expression by explicitly addressing the line and the entry names you want to change.

If any of the text entries could include spaces, then things could get trickier.