destring a complicated variable, with several nonnumerical forms

[Neg Kha]

Hi,

The variable grade in my dataset is badly collected and it has been saved as a string variable.
How can I destring it while some observations have % or not, U (for fails), commas and dots?

* Example generated by -dataex-. To install: ssc install dataex
clear
input str22 grade
"66,55%"
"81%"  
"72.5"  
"60%"  
"84,85%"
"84,4%"
"84,65%"
"81,35%"
"75%"  
"81,4%"
"79,5%"
"U"    
"81.7%"
"71,05%"
"91,45%"
"70%"  
"89,9%"
"83,8%"
"88,75%"
"82,65%"
"88,75"
"86,95%"
"90,35%"
"79.6"  
"74"
"82,6%"
"84,5%"
"77.6%"
"81"    
"67,5%"
"U"    

end

Thanks in advance

[Andrew Musau]

".a" identifies failures below. See

help missing

* Example generated by -dataex-. To install: ssc install dataex
clear
input str22 grade
"66,55%"
"81%"  
"72.5"  
"60%"  
"84,85%"
"84,4%"
"84,65%"
"81,35%"
"75%"  
"81,4%"
"79,5%"
"U"    
"81.7%"
"71,05%"
"91,45%"
"70%"  
"89,9%"
"83,8%"
"88,75%"
"82,65%"
"88,75"
"86,95%"
"90,35%"
"79.6"  
"74"
"82,6%"
"84,5%"
"77.6%"
"81"    
"67,5%"
"U"    

end

gen wanted= subinstr(subinstr(subinstr(grade, "%", "", .), ",", ".",.),"U", ".a",.)
destring wanted, replace

Res.:

. l, sep(0)

     +-----------------+
     |  grade   wanted |
     |-----------------|
  1. | 66,55%    66.55 |
  2. |    81%       81 |
  3. |   72.5     72.5 |
  4. |    60%       60 |
  5. | 84,85%    84.85 |
  6. |  84,4%     84.4 |
  7. | 84,65%    84.65 |
  8. | 81,35%    81.35 |
  9. |    75%       75 |
 10. |  81,4%     81.4 |
 11. |  79,5%     79.5 |
 12. |      U       .a |
 13. |  81.7%     81.7 |
 14. | 71,05%    71.05 |
 15. | 91,45%    91.45 |
 16. |    70%       70 |
 17. |  89,9%     89.9 |
 18. |  83,8%     83.8 |
 19. | 88,75%    88.75 |
 20. | 82,65%    82.65 |
 21. |  88,75    88.75 |
 22. | 86,95%    86.95 |
 23. | 90,35%    90.35 |
 24. |   79.6     79.6 |
 25. |     74       74 |
 26. |  82,6%     82.6 |
 27. |  84,5%     84.5 |
 28. |  77.6%     77.6 |
 29. |     81       81 |
 30. |  67,5%     67.5 |
 31. |      U       .a |
     +-----------------+

[Neg Kha]

".a" identifies failures below. See

help missing

* Example generated by -dataex-. To install: ssc install dataex
clear
input str22 grade
"66,55%"
"81%"
"72.5"
"60%"
"84,85%"
"84,4%"
"84,65%"
"81,35%"
"75%"
"81,4%"
"79,5%"
"U"
"81.7%"
"71,05%"
"91,45%"
"70%"
"89,9%"
"83,8%"
"88,75%"
"82,65%"
"88,75"
"86,95%"
"90,35%"
"79.6"
"74"
"82,6%"
"84,5%"
"77.6%"
"81"
"67,5%"
"U"

end

gen wanted= subinstr(subinstr(subinstr(grade, "%", "", .), ",", ".",.),"U", ".a",.)
destring wanted, replace

Res.:

. l, sep(0)

+-----------------+
| grade wanted |
|-----------------|
1. | 66,55% 66.55 |
2. | 81% 81 |
3. | 72.5 72.5 |
4. | 60% 60 |
5. | 84,85% 84.85 |
6. | 84,4% 84.4 |
7. | 84,65% 84.65 |
8. | 81,35% 81.35 |
9. | 75% 75 |
10. | 81,4% 81.4 |
11. | 79,5% 79.5 |
12. | U .a |
13. | 81.7% 81.7 |
14. | 71,05% 71.05 |
15. | 91,45% 91.45 |
16. | 70% 70 |
17. | 89,9% 89.9 |
18. | 83,8% 83.8 |
19. | 88,75% 88.75 |
20. | 82,65% 82.65 |
21. | 88,75 88.75 |
22. | 86,95% 86.95 |
23. | 90,35% 90.35 |
24. | 79.6 79.6 |
25. | 74 74 |
26. | 82,6% 82.6 |
27. | 84,5% 84.5 |
28. | 77.6% 77.6 |
29. | 81 81 |
30. | 67,5% 67.5 |
31. | U .a |
+-----------------+

This is great thanks! There are some observations that have quotations around the number like "76%". How can I get rid of those?
Edit: Actually, I figured it out. I can use `"" ' for that. Thanks anyways!!