Re: st: Mata order() indeterminate

[vwiggins@stata.com (Vince Wiggins, StataCorp)]

Brendan Halpin <brendan.halpin@ul.ie> asks about obtaining stable
ordering (sorts) when there are tied values and when using the
-order()- function in Mata.

Brendan provides an example, see below, where the ordering of ties is
done differently each time the functions are run.

As Joseph Coveney <stajc2@gmail.com> discusses, this randomness in
ordering tied values is intentional.  It keeps us from seeing
consistency where there is none.

That said, we have no problem with someone knowingly breaking ties
consistently.  A surefire solution is to create another column in your
ordering matrix that has a sequential ordering, then include that
column when using -order()-.  If you were ordering on the first column
of matrix y you would type -order(y, 1)-.  After inserting a second
column that is has sequential values you would type -order(y, 1..2)-.
That solution makes it explicit how you want your ties broken.

That said, there is another way.  Mata's -order()- and -sort()- use
Stata's internal sort seed, see -help sortseed-, to order ties
consistently.  Brendan can set that seed before using -order()- to get
reproducible sorts.  Just code

    : stata("set sortseed 12345")


We will consider exposing this approach in the documentation, and even
adding a -sortseed()- function to Mata.  There is, however, a clarity
in requiring that ties be broken explicitly.


 
-- Vince 
   vwiggins@stata.com

 
 
------------ Original message ---------------------------
 
From: brendan.halpin@ul.ie (Brendan Halpin)
Subject: st: Mata order() indeterminate
Sender: owner-statalist@hsphsun2.harvard.edu
Lines: 40

I'm checking a simulation in Mata, and find that setting the seed to the
same value does not yield the same results on repeated runs. I've
tracked it down to the use of -order()- to sort a matrix, where there
are many ties. It appears that -order()- brings in indeterminacy in
dealing with ties, but from somewhere other than the random-number system.

This snippet illustrates the issue:

mata:
x = range(1,10,1)
y = runiform(10,1):>0.5

for (i=1; i<=20; i++) {
  rseed(12345)
  x, x[order(y,1),]
}
end

Though x is unchanged, and the seed is set to the same value at each
pass, x[order(y,1)] changes. 

While this is disturbing, I presume it is consistent with Stata policy
with regard to sorting indeterminacy. 

How do I get repeatable sorting in this context?


Regards,

Brendan
-- 
Brendan Halpin, Head of Department, Sociology, University of Limerick, Ireland
Tel: w +353-61-213147  f +353-61-202569  h +353-61-338562;  Room F1-002 x 3147
mailto:brendan.halpin@ul.ie    ULSociology on Facebook: http://on.fb.me/fjIK9t
http://teaching.sociology.ul.ie/bhalpin/wordpress         twitter:@ULSociology
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/