vi/vim, how can I write out the nth word of all lines to a new file

Question

How can I write out the nth word of all lines to a new file? For example, I want the second word of all lines wrote out in new_file.txt:

old_file.txt:

pippo pluto paperino 
gigi lui marco 
piero andrea chiara

new_file.txt:

pluto  
lui   
andrea

Answer 1

First step, delete all but the second word on each line. This is a fairly straightforward :%norm command:

:%norm dwwhD

Which is like manually typing out dwwhD on each line, but faster. Second step, save to a new file:

:w new_file.txt

And then quit without saving so that old_file.txt stays unmodified:

:q!

Answer 2

:!cut % '-d ' -f2 > new_file.txt

Note that the default delimiter of cut is tab, so you'll want to use '-d ' to set the delimiter to space.

From inside of Vi you can use % as the replacement for the current filename.

Answer 3

You can do this from the terminal command line using a simple bash script. First loop through the contents of the file. Then for each line, pipe an echo into awk and have awk extract the second column. Lastly redirect that text into a new file.

Putting all of that together:

while read line; do
    echo $line | awk '{print $2}' >> new_file.txt
done <old_file.txt

Answer 4

Use a :s command with a regular expression:

:%s/\v^\s*\S+(\s+(\S+).*)?/\2

The \v tells Vim to use the "very magic" style of regexps, which is similar to extended regexps in grep, or pcre.

The regexp matches a single word at the start, possibly preceded by some optional whitespace. Then, after some more whitespace, the second word is captured inside group \2. Finally, a .* matches the whole rest of the line.

The part after the first word is optional (grouped in parentheses and made optional with the final ?). This ensures lines with a single word will still be matched, and in this case they'll be replaced with a blank line, which I imagine is what should be expected of a line with a single word, since the second work is blank in that case. This is one advantage of using regexps for this task, as you get more control over what to do with lines that have corner cases.

Finally, group \2 is used for replacement. It captured the second word on the line, by itself.

You can then save the modified contents to the new filename:

:saveas new_file.txt

This will preserve your unmodified old_file.txt and will make sure further modifications of the current buffer will only go to new_file.txt.