Conditionally delete xml tags

Question

In a file containing tags like these:

<div class="myclass">
Something irrelevant.
</div>
<div class="myclass">
MATCH ME
</div>

I would like to delete all tags not containing a specific pattern, here "MATCH ME" as an example. I can search for the opening of the tag, then do vat, inspect the selection, if it doesn't contain my pattern I hit d to delete it and move on to the next candidate with n, else I will hit n straight, continuing to the end of the file. Now I'd like to automate that, as I have a few files like that. How?

Answer 1

Assuming that MATCH ME is always a single line, you can use something like:

:g/^MATCH ME/-1,+1d

• First, :g/^MATCH ME/ will search for all instances of MATCH ME.
• We then run the :d (:delete) command on everything that matches with :g, but we also add a range of -1,+1, meaning one line before and one line after the matched line.

If your MATCH ME pattern may consist of multiple lines, we need a slightly more advanced version:

:g/^MATCH ME/call DeleteBetween('<div class="myclass">', '<\/div>')

Instead of adding a whole bunch on the command-line (which quickly becomes unreadable), I created a little helper function:

fun! DeleteBetween(start, end)
    " Reverse search for where we should start deleting
    execute '?' . a:start

    " Remember the line number
    let l:start = line('.')

    " Forward search for when we should stop deleting
    execute '/' . a:end

    " And delete the line range (:4,8d)
    execute ':' . l:start . ',' . line('.') . 'd'
endfun