Re: st: data management

[n j cox <n.j.cox@durham.ac.uk>]

Carole J. Wilson had a solution that solved the problem.
I agree with her analysis. You need to loop over observations.

Here is a slightly simpler solution along the same lines.

One way to tackle this is just to -count-. -count- is one of the most
neglected Stata commands, but see

SJ-7-1  pr0029 . . . . . . . . . . . .  Speaking Stata: Making it count
Q1/07   SJ 7(1):117--130                                 (no commands)
discusses count used with a loop over observations or variables

People who want to count often create a variable with 1s and 0s and then 
-summarize-. The sum of the values is, naturally, is the same as the 
number of 1s. If you do do this, make sure you do it directly: Use
-summarize, meanonly- and pick up -r(sum)-.

Using -egen, total()- is a similar method, but about 100 times as much
code for Stata to interpret. (-viewsource egen.ado- and -viewsource
_gtotal.ado- to see this.

-summarize- for counting can be a good method, but -count- is often better.

In the first example, Sim wants, I think,

. count if ID != 1 & size == 6
	& market_exit > market_entry[1]
	& market_entry < market_entry[1]

You can automate that:

gen exists_size_6 = 0

qui forval i = 1/`=_N' {
	count if ID != ID[`i'] & size == 6 & ///
		market_exit > market_entry[`i'] & ///
		market_entry < market_entry[`i']
	replace exists_size6 = r(N) in `i'
}

So, the basic idea is just

generate a count variable
loop over observations {
	count how many "match" this observation
	replace count variable with count in this observation
}

Nick
n.j.cox@durham.ac.uk

Sim Oertel

I have a question regarding the definition of a new variable in a data
set. I would like to define this new variable in the way that it will
tell me how many subjects with a specific characteristic exist at a
specific time. For example:  Suggesting a simplified data set of nine
subjects with information regarding their market-entry, market-exit, and
size. In this case, I would like to know how many subjects with the size
6 were in the market at the same time with each other subject.

Example:
ID            Market-Entry             Market-Exit         Size
 New_Variable
1             10                             98                       1
               1
2             15                             80                       1
               0
3             30                             110                     1
              2
4             41                             97                       1
               1
5             76                             85                       1
               1
6             77                            218                      2
              2
7             85                            220                      3
              2
8             82                            118                      6
              2
9             99                            218                      6
              2


For ID1, for example, subject 8 entered at time 82, while ID1
markets-exist is 98. The other subject with size 6 is ID9. But ID9
entered the market after the market-exist of ID1. So the variable should
hold the value 1 (One subject with size 6 was present in the market at
the same time as ID1). Regarding subject ID2, ID8 and ID 9 entered after
the market-exit of ID2, so the new variable should hold the value 0. For
ID3, the new variable should hold the value 2, because both subjects
with the size 6 were present in the market at the same time than ID3,
and so on.


*
*   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/