st: Short program to "collapse (# unique elements)": Use of nested loops and a "weights not allowed" message

["Chih-Mao Hsieh" <Hsieh@olin.wustl.edu>]

Hi statalisters,

I have been working on a short program that doesn't seem to work, I think I'm just missing a small mistake...  I have a data file with three columns: citing, cited, nclass.  For every "citing", there are multiple "cited", and for each "cited" there is a "nclass".  The file is sorted by citing, then nclass.  I need a program to count the number of unique "nclass" strings associated to each "citing".

As a simple example, given the following data file "data.dta":

citing     cited         nclass
100         20            12
100         22            15
100         23            15
101         32            14
101         33            15
101         34            15
101         40            17

I need the following output file:

citing    numpatclass
100            2             [12 and 15 are unique, 15 is repeated]
101            3             [14, 15, 17 are unique, 15 is repeated]

I have decided to do it by creating an intermediate file which I will later collapse(max):

citing     cited         nclass         indexpatclass
100         20            12                    1
100         22            15                    2
100         23            15                    2
101         32            14                    1
101         33            15                    2
101         34            15                    2
101         40            17                    3

"indexpatclass" indexes by 1 whenever a "citing" involves a new "nclass", and resets to 1 whenever a new "citing" begins.  So I have created a short program.  It sorts by "citing" and "nclass", then it uses a while-loop, and then two if-loops.  But there are two problems: (1) I am getting a "weights not allowed" message when I try to run it.  (2) I am also not sure whether I am properly nesting my loops.  Can anybody provide any insight?  Or alternatively, is there a much simpler way to do what I am attempting?

Thanks, --Chihmao.

--------------------------------

# delimit cr
program define uniqpatclass
use c:\temp\data
generate indexpatclass=0
sort citing nclass
replace indexpatclass=1 in 1
generate id=_n

while id<_N {
   if citing[_n]==citing[_n-1] {
      if nclass[_n]==nclass[_n-1] {
         replace indexpatclass[_n]=indexpatclass[_n-1]
         id = `id' + 1
      }
      else {
   replace indexpatclass[_n]=indexpatclass[_n-1]+1
   id = `id' + 1
   }
   }
   else {
replace indexpatclass[_n]=1}
id = `id' + 1
}
end



*
*   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/