replace subsequent values

[Qian Cai]

Hi Stata experts,

below is my dataset, what I am looking forward to is by ID, if duration=1 & duration[_n+1] ==0 & duration[_n+2] ==0, then replace the duration[_n+1] ==1 & duration[_n+2] ==1. Can you help me out of this? Thank you very much.

clear
input float(ID duration var3)
1 30 1
1 30 0
1 20 0
1 16 0
2 35 0
2 86 1
2 24 0
2 87 0
3 23 0
3 86 1
3 60 1
3 87 1
3 34 0
3 14 1
3 67 1
3 89 0
3 43 0
4 32 1
4 54 1
4 100 1
4 58 1
4 85 0
4 34 0
4 16 0
end

blow dataset is what I expect

clear
input float(ID duration var3)
1 30 1
1 30 1
1 20 1
1 16 0
2 35 0
2 86 1
2 24 1
2 87 1
3 23 0
3 86 1
3 60 1
3 87 1
3 34 0
3 14 1
3 67 1
3 89 1
3 43 1
4 32 1
4 54 1
4 100 1
4 58 1
4 85 1
4 34 1
4 16 0
end

[Qian Cai]

sorry, just to be clear, I only hope to replace the subsequent two 0(s) with 1 if period[_n]=1. Look forward to your guidance.

[Andrew Musau]

* Example generated by -dataex-. For more info, type help dataex
clear
input float(ID duration var3)
1  30 1
1  30 0
1  20 0
1  16 0
2  35 0
2  86 1
2  24 0
2  87 0
3  23 0
3  86 1
3  60 1
3  87 1
3  34 0
3  14 1
3  67 1
3  89 0
3  43 0
4  32 1
4  54 1
4 100 1
4  58 1
4  85 0
4  34 0
4  16 0
end

gen long obsno=_n
bys ID (obsno): gen tag= (var3[_n-2] & !var3[_n-1])|(var3[_n-1] & !var3[_n+1])
replace var3= tag if !var3 & tag

Res.:

. l, sepby(ID)

     +------------------------------------+
     | ID   duration   var3   obsno   tag |
     |------------------------------------|
  1. |  1         30      1       1     1 |
  2. |  1         30      1       2     1 |
  3. |  1         20      1       3     1 |
  4. |  1         16      0       4     0 |
     |------------------------------------|
  5. |  2         35      0       5     0 |
  6. |  2         86      1       6     1 |
  7. |  2         24      1       7     1 |
  8. |  2         87      1       8     1 |
     |------------------------------------|
  9. |  3         23      0       9     0 |
 10. |  3         86      1      10     1 |
 11. |  3         60      1      11     0 |
 12. |  3         87      1      12     1 |
 13. |  3         34      0      13     0 |
 14. |  3         14      1      14     1 |
 15. |  3         67      1      15     1 |
 16. |  3         89      1      16     1 |
 17. |  3         43      1      17     1 |
     |------------------------------------|
 18. |  4         32      1      18     0 |
 19. |  4         54      1      19     0 |
 20. |  4        100      1      20     0 |
 21. |  4         58      1      21     1 |
 22. |  4         85      1      22     1 |
 23. |  4         34      1      23     1 |
 24. |  4         16      0      24     0 |
     +------------------------------------+

[Nick Cox]

Another way to do it.

clear
input float(ID duration var3)
1 30 1
1 30 0
1 20 0
1 16 0
2 35 0
2 86 1
2 24 0
2 87 0
3 23 0
3 86 1
3 60 1
3 87 1
3 34 0
3 14 1
3 67 1
3 89 0
3 43 0
4 32 1
4 54 1
4 100 1
4 58 1
4 85 0
4 34 0
4 16 0
end

gen long pstime = _n 

gen when1 = pstime if var3 == 1 

bysort ID (pstime) : replace when = when[_n -1] if missing(when)

clonevar wanted = var3 

replace wanted = 1 if wanted == 0 & inlist(pstime - when, 1, 2)

list, sepby(ID)


     +------------------------------------------------+
     | ID   duration   var3   pstime   when1   wanted |
     |------------------------------------------------|
  1. |  1         30      1        1       1        1 |
  2. |  1         30      0        2       1        1 |
  3. |  1         20      0        3       1        1 |
  4. |  1         16      0        4       1        0 |
     |------------------------------------------------|
  5. |  2         35      0        5       .        0 |
  6. |  2         86      1        6       6        1 |
  7. |  2         24      0        7       6        1 |
  8. |  2         87      0        8       6        1 |
     |------------------------------------------------|
  9. |  3         23      0        9       .        0 |
 10. |  3         86      1       10      10        1 |
 11. |  3         60      1       11      11        1 |
 12. |  3         87      1       12      12        1 |
 13. |  3         34      0       13      12        1 |
 14. |  3         14      1       14      14        1 |
 15. |  3         67      1       15      15        1 |
 16. |  3         89      0       16      15        1 |
 17. |  3         43      0       17      15        1 |
     |------------------------------------------------|
 18. |  4         32      1       18      18        1 |
 19. |  4         54      1       19      19        1 |
 20. |  4        100      1       20      20        1 |
 21. |  4         58      1       21      21        1 |
 22. |  4         85      0       22      21        1 |
 23. |  4         34      0       23      21        1 |
 24. |  4         16      0       24      21        0 |
     +------------------------------------------------+

See also https://journals.sagepub.com/doi/pdf...6867X231196519 including Section 5.2.