> I want to check several new-articles on the occurences of words
> Therefore the two questions:
>
> 1) How to let vim place a carriage return after each word in an news-article?
You can use
:%s/\>/\r/g
though other non-word characters (defined by the 'iskeyword'
setting) will end up leading lines.
> And 2) If you have a text like this
>
> AAA
> AAA
> AAA
> BBB
> CCC
> CCC
> CCC
> CCC
>
> can you let vim count the occurences of these words and add a number
> before them and delete all but the first one?
Have you already normalized the case of the words?
> So that the text above becomes:
>
> 3 AAA
> 1 BBB
> 4 CCC
You might do a multi-step process such as:
:%s/\>/\r/g " break after all words
:%s/\W//g " remove all non-word characters
:v/\w/d " delete any lines that end up blank
:%s/.*/\U&/ " uppercase the entire document to normalize
:%sort " sort the results as required by "uniq"
:%!uniq -c " use the *nix "uniq" tool to prepend counts
It gets a bit hairier if you need to do this on a non-*nix box
where "uniq" isn't available. You might be able to replace the
"uniq" command with something like the following in pure Vim
(requires vim7, as the Dict type wasn't introduced until vim7)
:let d={} |g/^/let d[getline('.')]=get(d, getline('.'), 0)+1
:%s/^\(.*\)\(\n\1\)*/\=d[submatch(1)]."\t".submatch(1)
:unlet d
which builds a dictionary ("d") mapping word->count (the ":g"
command) and then collapses multiple repeats into a single item
prefixed by the count for that item (the ":%s" command), then
ditches ("unlet"s) the dictionary to free up memory.
Hope this helps,
-tim
--~--~---------~--~----~------------~-------~--~----~
You received this message from the "vim_use" maillist.
For more information, visit http://www.vim.org/maillist.php
-~----------~----~----~----~------~----~------~--~---
