Sort country identifier according to ranking

[Florian Schoner]

Dear all,
I have the following data. I want to replace the entries of rank_y_628 by the country identifiers (icode_imf) according to the ranking given by rank_y_628.

     +------------------------------+
     | icode_~f   rank~628   obsnum |
     |------------------------------|
  1. |      111          9        1 |
  2. |      112         16        2 |
  3. |      122         31        3 |
  4. |      124         29        4 |
  5. |      128         20        5 |
     |------------------------------|
  6. |      132         25        6 |
  7. |      134         12        7 |
  8. |      138         33        8 |
  9. |      142         22        9 |
 10. |      144         32       10 |
     |------------------------------|
 11. |      146         34       11 |
 12. |      156         28       12 |
 13. |      158          2       13 |
 14. |      172         17       14 |
 15. |      174         18       15 |
     |------------------------------|
 16. |      178         30       16 |
 17. |      182         35       17 |
 18. |      184         21       18 |
 19. |      186          6       19 |
 20. |      193         14       20 |
     |------------------------------|
 21. |      196          8       21 |
 22. |      199          5       22 |
 23. |      273         26       23 |
 24. |      528          3       24 |
 25. |      532          4       25 |
     |------------------------------|
 26. |      534          1       26 |
 27. |      536         19       27 |
 28. |      542          7       28 |
 29. |      548         27       29 |
 30. |      566         24       30 |
     |------------------------------|
 31. |      576         11       31 |
 32. |      578         10       32 |
 33. |      935         15       33 |
 34. |      944         23       34 |
 35. |      964         13       35 |

I tried to achieve that with the following code.

local count=_N
forval i=1/`count'{
local hold=rank_y_628[`i']
replace rank_y_628=icode_imf[`i'] if `hold'==obsnum
}

Somehow though, this does not work for all observations. More specifically, I receive the following output:

     +----------+
     | rank~628 |
     |----------|
  1. |        9 |
  2. |      158 |
  3. |      528 |
  4. |       29 |
  5. |      199 |
     |----------|
  6. |      186 |
  7. |      542 |
  8. |      196 |
  9. |      111 |
 10. |       32 |
     |----------|
 11. |       34 |
 12. |      134 |
 13. |      964 |
 14. |       17 |
 15. |       18 |
     |----------|
 16. |      112 |
 17. |      172 |
 18. |      174 |
 19. |      536 |
 20. |      128 |
     |----------|
 21. |        8 |
 22. |        5 |
 23. |       26 |
 24. |      566 |
 25. |      132 |
     |----------|
 26. |      273 |
 27. |       19 |
 28. |        7 |
 29. |      124 |
 30. |       24 |
     |----------|
 31. |      122 |
 32. |      144 |
 33. |      138 |
 34. |      146 |
 35. |       13 |

I'd be grateful for an explanation what went wrong and how I could fix this issue.

Kind regards,
Florian

P.S. Unfortunately I'm not able to use dataex since I'm using Stata 15.0 and cannot access user-written commands.