How to exclude . in the count of unique values?

[Sehajroop Bath]

I'm trying to follow what they have listed on the guide here but they don't include the case where one might want to exclude . in their count of unique values.

[Rich Goldstein]

use the user-written -distinct- command (by the authors of the FAQ you link to); use -findit- or -search- to find and download; see (after installing)

help distinct

[Nick Cox]

There are various way to tackle that, not least specifying if !missing() with a variable.

Here's another: distinct from the Stata Journal excludes missings by default. Consider this example:

. sysuse auto, clear
(1978 Automobile Data)

. distinct rep78

------------------------------
       |     total   distinct
-------+----------------------
 rep78 |        69          5
------------------------------

. distinct rep78, missing

------------------------------
       |     total   distinct
-------+----------------------
 rep78 |        74          6
------------------------------

Incidentally, an update of distinct in press for Stata Journal 20(4) has sorting options, such as

 
. distinct, sort(distinct)

-------------------------------------
              |     total   distinct
--------------+----------------------
      foreign |        74          2
        rep78 |        69          5
     headroom |        74          8
         turn |        74         18
        trunk |        74         18
          mpg |        74         21
 displacement |        74         31
   gear_ratio |        74         36
       length |        74         47
       weight |        74         64
         make |        74         74
        price |        74         74
-------------------------------------

. distinct, sort(distinct descending)

-------------------------------------
              |     total   distinct
--------------+----------------------
        price |        74         74
         make |        74         74
       weight |        74         64
       length |        74         47
   gear_ratio |        74         36
 displacement |        74         31
          mpg |        74         21
        trunk |        74         18
         turn |        74         18
     headroom |        74          8
        rep78 |        69          5
      foreign |        74          2
-------------------------------------

.

[Sehajroop Bath]

How do I store it in a variable? STATA is giving me an error when I try to do that.

[Rich Goldstein]

you don't say what you tried so we can't say what you did incorrectly; I do note that the result is saved (see he help file); I also note that you probably do not really want a variable that is constant for each and every observation so what do you want? it can be saved as a local; if you really want it as a variable:

gen junk=r(ndistinct)

replace "junk" with whatever you want your new variable to be called

save as a local:

local junk=r(ndistinct)

[Sehajroop Bath]

you don't say what you tried so we can't say what you did incorrectly

This is the error I'm getting in the debugging log:

- gen nvals = `this_variable'(ndistinct)
= gen nvals = rsex(ndistinct)
unknown function rsex()

[Nick Cox]

That is not what Rich Goldstein is suggesting. I think we need to know more of what you are trying to do to give further advice.