Question

I have an XML file where I only care about the size attribute of a certain element.
First I used

global!/<proto name="geninfo"/d

to delete all lines that I don’t care about. That leaves me a whole bunch of lines that look like this:

<proto name="geninfo" pos="0" showname="General information" size="174">

I want to delete everything but the value for “size.”
My plan was to use substitute to get rid of everything not matching ‘size=”[digit]”‘, the remove the string ‘size’ and the quotes but I can’t figure out how to substitute the negation of a string.

Any idea how to do it, or ideas on a better way to achieve this? Basically I want to end up with a file with one number (the size) per line.

Answer 1

You can use matching groups:

:%s/^.*size="\([0-9]*\)".*$/\1/

This will replace lines that contain size="N" by just N and not touch other lines.

Explanation: this will look for a line that contains some random characters, then somewhere the chain size=", then digits, then ", then some more random characters, then the end of the line. Now what I did is that I wrapped the digits in (escaped) parenthesis. That creates a group. In the second part of the search-and-replace command, I essentially say “I want to replace the whole line with just the contents of that first group” (referred to as \1).