|
|  |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: Re: nearmrg
"Inamo, Jocelyn F. M.D., Ph.D." <[email protected]> wrote:
> I have some trouble using the nearmrg command written by M. Blasnik.
Let's say that I have two files ageinfo1 and ageinfo2, and that I want to
compare income in a matched subgroup. After doing the commands below, I
have a message error saying "using data not sorted"
use ageinfo2
nearmrg gender using ageinfo1, nearvar(age) genmatch(_match)
In an other set of experiment I had an other message " gender age" not
unique, although I used a random process to build age, so that any
combination based on age should be unique.
What step did I miss? I know this problem was pointed out, and fixed, a
couple of weeks ago. But it simply does not work in my hands.
Thanks
<snip>
In answer to your first question, I believe nearmrg, like merge, requires
that the -using- dataset be sorted by the variables. In your example,
ageinfo1 should be sorted by gender age. Try:
use ageinfo1
sort gender age
save, replace
use ageinfo2
nearmrg gender using ageinfo1, nearvar(age) genmatch(_match)
In your other example, I'm not sure how you randomly generated ages, but if
they were integers, there is no reason to think they would have to be
unique. -nearmrg- checks that the values uniquely identify the observations
using Stata's -isid- command.
I did find at least one strange piece of code looking at nearmrg.ado (it
saves a file that isn't a temporary file, needlessly). I may fix it soon
and send an update to SSC.
Michael Blasnik
[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/
