Highlight columns which differer from previous lines

I have text file with numerical data written i开发者_如何学Pythonn lines; there are interspersed input lines (unknowns) and residuals lines (which are supposed to be minimized). I am investigating ways how the iterative solver handles various cases, and would like to highlight every (space-delimited) field in the residuals line which is (textually) different from the same field in the previous residuals line (2 lines above, better given by a regexp). I am free to decorate beginnings of the lines as I like, if that helps.

Is this at all possible with Vim and some regexp magic?

Example file:

input 1 2 3 4 5 6
errors .2 .2 .3 .1 0 0
input 1 2.1 2.9 4 5 6    ## here, 2.1 and 2.9 should be highlighted
errors .21 .3 .44 .3 0 0
input 1 2 3 3.9 5.2 6    ## here, 2, 3, 3.9 and 5.2 should be highlighted
errors .2 .2 .34 .9 1 0

Note: I could code script extracting differences in Python, but I want to have a look at both the actual data and the changes. When it does not work with Vim, I will process it with python and output highlighted HTML, but I will lose automatic folding capabilities.

I think that if you use arrays rather than a regex for the comparison, it might be a lot easier:

<?php
$lastRow=array()
$h=fopen('file','r');
while ($r=fgetcsv($h,0,' ')) // Retrieve a line into an array and split on spaces
{
    if (count($lastRow)==0)
    {
        $lastRow=$r; // Store last line
    }

    $count++;

    if ($r[0]=='input')
    {
        /*
         * this won't find any differences the first run through
         */
        $diffs=array_diff($lastRow,$r); // Anything that has changed from the last "input" line is now in $diffs
        $lastRow=$r; // Only copy into $lastRow if it's an "input" line
    }
    /*
     *  Put your code to output line here with relevant CSS for highlighting
     */
}
fclose($h);

?>

I've not tested this code but I think it's shows how you could get a solution without delving into regex