keeping observations with greatest value within each group

[Sujin Shin]

Hello, I am struggling with organising the following data set.
[CODE]

* Example generated by -dataex-. For more info, type help dataex
clear
input long h_pid float(h_pid1 nhpid age)
1502 1503 16 34
1502 1503 16 40
1502 1503 16 33
1502 1503 16 35
1502 1503 16 41
1502 1503 16 32
1502 1503 16 38
1502 1503 16 43
1502 1503 16 37
1502 1503 16 45
1502 1503 16 36
1502 1503 16 42
1502 1503 16 30
1502 1503 16 39
1502 1503 16 31
1502 1503 16 44
1502 1504 12 45
1502 1504 12 37
1502 1504 12 38
1502 1504 12 35
1502 1504 12 44
1502 1504 12 36
1502 1504 12 34
1502 1504 12 40
1502 1504 12 42
1502 1504 12 41
1502 1504 12 43
1502 1504 12 39
3802 3803  7 42
3802 3803  7 34
3802 3803  7 39
3802 3803  7 43
3802 3803  7 44
3802 3803  7 37
3802 3803  7 36
3802 3804  7 46
3802 3804  7 35
3802 3804  7 45
3802 3804  7 41
3802 3804  7 40
3802 3804  7 38
3802 3804  7 33
4502 4503 16 33
4502 4503 16 41
4502 4503 16 38
4502 4503 16 42
4502 4503 16 46
4502 4503 16 43
4502 4503 16 39
4502 4503 16 34
4502 4503 16 36
4502 4503 16 47
4502 4503 16 45
4502 4503 16 44
4502 4503 16 37
4502 4503 16 32
4502 4503 16 35
4502 4503 16 40
end

I want to keep the h_pid1 with the greatest nhpid values within the h_pid group. For example, within the group, h_pid==1502, I only want to keep the observations of h_pid1==1503 because it have the highest nhpid value among the same h_pid group.

I tried using bysort to start but I struggled to get the result that I want.

It would be appreciated if you could help me with this. Thank you

[Carlo Lazzaro]

Sujin:
you may want to try:

. bysort h_pid: egen wanted=max( h_pid1 )

[Nick Cox]

bysort h_pid (nhpid)  : keep if h_pid1 == h_pid1[_N]

list, sepby(h_pid)

     +------------------------------+
     | h_pid   h_pid1   nhpid   age |
     |------------------------------|
  1. |  1502     1503      16    41 |
  2. |  1502     1503      16    33 |
  3. |  1502     1503      16    40 |
  4. |  1502     1503      16    43 |
  5. |  1502     1503      16    35 |
  6. |  1502     1503      16    45 |
  7. |  1502     1503      16    36 |
  8. |  1502     1503      16    37 |
  9. |  1502     1503      16    39 |
 10. |  1502     1503      16    30 |
 11. |  1502     1503      16    42 |
 12. |  1502     1503      16    32 |
 13. |  1502     1503      16    34 |
 14. |  1502     1503      16    38 |
 15. |  1502     1503      16    44 |
 16. |  1502     1503      16    31 |
     |------------------------------|
 17. |  3802     3804       7    46 |
 18. |  3802     3804       7    40 |
 19. |  3802     3804       7    35 |
 20. |  3802     3804       7    45 |
 21. |  3802     3804       7    41 |
 22. |  3802     3804       7    33 |
 23. |  3802     3804       7    38 |
     |------------------------------|
 24. |  4502     4503      16    46 |
 25. |  4502     4503      16    41 |
 26. |  4502     4503      16    32 |
 27. |  4502     4503      16    34 |
 28. |  4502     4503      16    37 |
 29. |  4502     4503      16    43 |
 30. |  4502     4503      16    33 |
 31. |  4502     4503      16    40 |
 32. |  4502     4503      16    44 |
 33. |  4502     4503      16    42 |
 34. |  4502     4503      16    45 |
 35. |  4502     4503      16    35 |
 36. |  4502     4503      16    38 |
 37. |  4502     4503      16    47 |
 38. |  4502     4503      16    36 |
 39. |  4502     4503      16    39 |
     +------------------------------+