[patch #5402] RANK draft 1

[John Darrington]

Follow-up Comment #15, patch #5402 (project pspp):

> [...] if you have multiple discontiguous SPLIT FILE groups with the same
split values, then they will all be ranked together as one big group, but
SPSS ranks each group separately. I had code that fixed this, but, um, it
didn't work due to a basic misconception of mine. It should be easy to fix,
but I don't have time tonight.

What was your solution to this?  My idea would be to add a new variable and a
transformation which populates it with unique group values. Eg:

G   $G_
20   1
20   1
30   2
30   2
30   2
20   3
20   3
30   4

Then we sort by $G_ instead of G.  This wouldn't add too much extra overhead,
since we have one transformation anyway.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/patch/?5402>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/