CODEWITHSUNDEEP
vim
Darius Schindler
Darius Schindler

Reputation: 33

Split a file to multiple files automaticaly in VIM?

I'm new in VIM and I try to split a file in multiple files.This is a test file:

Something1;XXXword;blabla(about 500 signs);
Something1;XXXword;(about 500 signs);
Something1;YYYword;(about 500 signs);
Something1;RRRword;(about 500 signs);

XXX could be a word 2-20 characters long. When the following word (XXX/YYY/RRR) changes then before "Something1" should be a cut and the following lines till XXX changes should be an another new file and so on.

It should be so:

File1:
Something1;XXXword;blabla(about 500 signs);Something1;XXXword;(about 500 signs);

File2:
Something1;YYYword;(about 500 signs);

File3:
Something1;RRRword;(about 500 signs)

Is there a way to do this like a pro? Thanks :)

Upvotes: 0

Views: 457

Answers (2)

Peter Rincker
Peter Rincker

Reputation: 45097

I would recommend a different tool, like Awk.

awk -F';' '{printf "%s", $0 >> $2}' your_file.txt

This will split each line into columns separated by ;. Each line will be appended (>>) to a file named after the 2nd column, $2 (e.g. XXXword). Append/print the whole line, $0, except the newline (printf "%s") to the new file so everything is one long line.

Note: I am using gawk as my awk implementation, you may need to make adjustments depending on your awk implementation.

For the more complex scenario

In the following case where you had XXX, YYY, XXX:

Something1;XXXword;blabla(about 500 signs);
Something1;YYYword;(about 500 signs);
Something1;XXXword;(about 500 signs);

If this should yield 3 files (1 YYY file and 2 XXX files) then we can use Awk as well:

awk -F';' 'last != $2 {f[$2]++} {printf "%s", $0 >> $2 f[$2]; last = $2}' your_file.txt

This will yield files: XXXword1, XXXword2, and YYYword1

This is the similar to the awk example above except we use a dictionary/array to store the number of times the 2nd column changes, f[$2]++, from the previous line last != $2 {...}. Making sure to set last to the 2nd column after printing each line. Output the line, $0, to a file named after $2 f[$2] (adjacent variable and string will be concatenated).

Upvotes: 6

Luc Hermitte
Luc Hermitte

Reputation: 32926

You'll have to program it as you would have programmed it in any other language. My first reflex would have been Perl BTW.

function! s:split(root) abort
  " todo: check empty buffers
  let lines = getline(1, '$')
  let nb_lines = len(lines)
  let files = []

  let crt = 0
  while crt < nb_lines
    " I suppose the word is the second field in a .csv file
    let word = matchstr(lines[crt], '^[^;]*;\zs[^;]*\ze;')
    " This is where the real magic happens, see :h /\@!
    let next = match(lines, '^[^;]*;\(\('.word.'\)\@![^;]\)*;', crt)
    if next == -1 | let next = nb_lines | endif
    let files += [ lines[crt : (next-1)] ]
    let crt = next
  endwhile

  echo files
endfunction

command! -nargs=1 SplitBuffer :call s:split("<args>")

Instead of let files += [ something ], you will want to execute

:let index = 0
...
:for...
    ...
    :call writefile(a:root.index, lines[crt : (next-1)])
    :let index += 1
:endfor

EDIT:

In case the sequence XXX, YYY, XXX shall lead to two files instead of 3, it can be done with this (convoluted and untested) oneliner -- still, prefer @Peter Rincker's awk based solution.

:call map(getline(1, '$'), 'writefile(v:val, split(v:val, ";")[1], "a")')

Upvotes: 1

Related Questions