Bashscripts.org

Hi,

I have a sample text file:



New lines in the file may or may not occur.

I would like to get only those parts of the file which are between the closest 'category' tags, so in this example:



I am trying to force awk to do that like that:



But this command gives me only:



Could anyone point me how to write the command properly?

Regards,
Robert

Hi Robert, and welcome!

Try something like this:



I ran this against your sample text and got:



I hope this helps!
-J

So, a further explanation of the sed expression posted earlier:

We're using sed's search and print functions, and capturing the data that we want using parentheses ().

sed -n
-> This tells sed to run 'silently', only printing what we specify with the 'p' command

s/
-> This begins our search string

^.*<category.name="[^"]*">
-> This part of the regex matches anything starting at the beginning of the line, up through the opening 'category' tag

\([^<]*\).*
-> This groups together everything after the opening category tag that *doesn't* match the '<' character, which should mark the beginning of the closing </category> tag

/\1/
This says that we want to work with the information within that set of parens. If there were multiple sets of parentheses, the next set would be \2, then \3 and so on.

p
The "p" command tells sed we want to go ahead and print the matches.

I hope this helps!
-J

Hi Jeo,

Thanks for your all-embracing answer. You are right - it was my mistake that I have cut my sample file too much :/

Imagine a slightly more complex file:



I would like to get:



from it. Notice that I would like to have my 'category' tags included in my result.

Could you tell me how to rewrite the command?

There are a few ways to do this actually! Let's start with the sed command we already used above, but instead of using (parentheses) to isolate what's between the tags, let's just print the tags and everything between!



This will break if the category tags are on separate lines though.

Jeo,

I really appreciate your help. I have one more ask.

What about if I could not say are there any newline's characters in my file or not? So my file can be like that:

I'd use Perl and some of the HTML parsers written for it.