Changing Values for IDs

[Simeon Lauterbach]

Dear all,

I am using Stata 16.1. Unforunately the dataex command does not give me what I want. I have a rather simple question, but I do not how to code it. As you can see from the picture above, I would like that the values of pid_7 take the values of pid_6. For example, in the two cases I highlighted, I would wish that the values 2 of PID_7 change to 3 (pid_6) or the values 4 change to 5. Anyone has an idea of how to code that?

Best,

Simeon

[Ken Chui]

Unforunately the dataex command does not give me what I want.

Try this:

dataex round parish_name hhid_6 pid_6 hhid_7 pid_7 pid_code

or this to get into these two specific HHID:

dataex round parish_name hhid_6 pid_6 hhid_7 pid_7 pid_code if inlist(hhid_6, "H2810501", "H2810601")

Also, your scheme may need a bit more explanation. For example, how would you resolve the first 4 cases? If the two 7's are changed to 8 (because pid_6 is 8), what would happen to cases 3 and 4 where pid_7 is already 8? In some way, I think it may make more sense to regroup them and reindex them, like this:

clear
input str3 (x1 x2)
A a1
A a1
A a2
A a2
A a2
B b3
B b3
B b4
B b5
B b5
end

egen newvar = group(x1 x2)
list, sep(0)

Results:

     +------------------+
     | x1   x2   newvar |
     |------------------|
  1. |  A   a1        1 |
  2. |  A   a1        1 |
  3. |  A   a2        2 |
  4. |  A   a2        2 |
  5. |  A   a2        2 |
  6. |  B   b3        3 |
  7. |  B   b3        3 |
  8. |  B   b4        4 |
  9. |  B   b5        5 |
 10. |  B   b5        5 |
     +------------------+

Your problem can also be caused by something upstream. Without knowing how they were merged and how those pid were computed, it's hard to tell/help.