About Foreach or Forvalue

[Esther Chau]

Hi! I am trying to create seven summary variables named den_1 to den_7 to simplify the results I have from 34 variables named total1-total34. Although there are 34 different groups, each observation actually only belongs to 7 groups at most. The number of groups that each belongs is shown by the variable group_n.

Enquiry:
Is possible that the -foreach- command or -forvalue- command can search over the 34 variables, take out the values, and put it in a new sets of variables (den_1- den_7)?


Data example skipping variables for simplification:

* Example generated by -dataex-. For more info, type help dataex
clear
input float(total1 total2 total3 total32 total33 total34 group_n)
208 207  . . . . 4
208 207  . . . . 4
  .   .  . . . . 0
  .   .  . . . . 1
208 207  . . . . 4
  .   .  . . . . 1
  .   .  . . . . 4
  .   .  . . . . 0
  .   .  . . . . 2
  .   .  . . . . 1
  .   .  . . . . 0
  .   .  . . . . 2
  .   .  . . . . 0
  .   .  . . . . 1
  .   .  . . . . 0
208 207  . . . . 4
  .   .  . . . . 2
  .   .  . . . . 3
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 2
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . 5 . 2
208 207  . . . . 5
  .   .  . . . . 0
  .   .  . . . . 3
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 2
  .   .  . . . . 2
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 2
  .   .  . . . . 0
208 207 50 . . . 5
  .   .  . . . . 1
  .   .  . . . . 3
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 0
208 207  . . . . 5
  .   .  . . . . 2
  .   .  . . . . 2
  .   .  . . . . 1
  .   .  . . . . 1
  .   .  . . . . 2
208 207  . . . . 5
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 1
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 3
  .   .  . . . . 2
  .   .  . . . . 0
  .   .  . . . . 2
  .   .  . . . . 0
208 207  . . . . 5
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 1
  .   .  . . . . 2
208 207  . . . . 6
  .   .  . . . . 4
  .   .  . . . . 2
208 207  . . . . 6
  .   .  . . . . 2
  .   .  . . . . 2
  .   .  . . . . 3
208 207  . . . . 5
  .   .  . . . . 3
  .   . 50 . . . 2
  .   .  . . . . 0
  .   .  . . . . 2
  .   .  . . . . 0
  .   .  . . . . 3
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 0
  .   . 50 . . . 2
  .   . 50 . . . 3
  .   .  . . . . 0
  .   .  . . . . 3
  .   .  . . . . 0
  .   .  . . . . 0
208 207  . . . . 5
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 2
208 207  . . . . 4
  .   .  . 5 . . 3
  .   .  . . . . 2
  .   .  . . . . 3
  .   .  . . . . 2
  .   .  . . . . 5
  .   .  . . . . 0
  .   .  . . . . 0
  .   .  . . . . 0
end

Thank you all very much!!

[Carlo Lazzaro]

Esther:
assuming that you're interested in the mean of those variables, you may want to consider something along the following lines (groups from 0 to 7, bounds included):

. forvalues i = 0(1)7 {
  2. egen wanted`i'=rowmean(total*)
  3.  }

Obvuously, what above can be tweaked according to your reseach goals, if different.

[Esther Chau]

Thank you for your reply, Carlo! Is there a way to just copy the numbers to one of the new variables?
As in, if searching isn't a problem, I could put

 replace wanted1 = total1
replace wanted2 = total4

[Carlo Lazzaro]

Esther:
assuming that you're interested in the mean of those variables, you may want to consider something along the following lines (groups from 1 to 7, bounds included):

. forvalues i = 1(1)7 {
  2. egen wanted`i'=rowmean(total*) if group==`i'
  3. }

What above amends my previous code and, hoperfully, could save your time.

[Nick Cox]

I don't see any immediate point in creating 7 new variables.

The results for different groups remain separate with

egen rowmean = rowmean(total*) 

and if later you really need distinct new variables you can fire up separate.

[Esther Chau]

Thank you very much indeed for your help, Nick Cox and Carlo Lazzaro!!