Thanks to Kit Baum, a revised version of -rowsort-
is now available from SSC.
Recently Alice Dobson posted a problem for which
I recommended -rowsort- from SSC. Alice also
pointed out that in a problem such as hers in
which missing values are present -rowsort- does
not operate as desired.
Suppose in a one observation (row) of a set of
variables the values are 1 . 3 . 5 . 7.
-rowsort- by default follows standard Stata
rules and orders these 1 3 5 7 . . . and with
a -descend- option it orders them . . . 7 5 3 1.
The latter then makes it difficult to read off
largest, second largest, and so forth, as might
often be wished.
An easy work-around is to hit the variables
concerned with -mvencode- and then adjust for what
you did.
However, as earlier indicated, I have now added
an option that specifies that missings should be
treated as some specified number when sorting is
being done. Normally, but not necessarily, the
user would specify a large negative integer lower
than the minimum in the data. It would be possible
to do this automatically, but giving the user
this handle adds some further control.
-rowsort- has been revised at the same time and
now requires Stata 8. (In some future revision
using Mata for the sorting is to be considered.)
The previous -rowsort- has now been renamed -rowsort6-
and will work with Stata 6 or 7, but it has not
otherwise been changed (i.e. it does not have this
new option).
Thanks to Alice for her interesting problem.
Nick
[email protected]
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
