Splitting values in Stata and save them in a new variable

Question

I have a numeric variable with values similar to the following system

1
2
12
21
2

I would like to split the values which have length > 1 and put the second half of the value in another variable.

So the second variable would have the values:

.
.
2
1
.

Theoretically I would just use a simple replace statement, but I am looking for a code/loop, which would recognize the double digit values and split them automatically and save them in the second variable. Because with time, there will be more observations added and I cannot do this task manually for >10k cases.

Answer 1

Here's one approach:

clear 
input foo 
1
2
12
21
2
end 

generate foo1 = floor(foo/10)
generate foo2 = mod(foo, 10)

list 

     +-------------------+
     | foo   foo1   foo2 |
     |-------------------|
  1. |   1      0      1 |
  2. |   2      0      2 |
  3. |  12      1      2 |
  4. |  21      2      1 |
  5. |   2      0      2 |
     +-------------------+

More on these functions here, here and here.

If zeros for the first part should be missing, then

replace foo1 = . if foo1 == 0 

or (to do it in one)

generate foo1 = floor(foo/10) if foo >= 10 

The code is also good for any arguments with three digits or more.