Stata: Systematically replace characters in a string variable

28,002

The substr() function (not substring(); not a command) is not as helpful here as its sibling, subinstr(). Documented in the same place: start at help functions.

. clear 
. input str7 myvar 

         myvar
1. CC547A1
2. VC549F
3. PC5297
4. end 

. replace myvar = subinstr(myvar, "CC", "18", .) 
(1 real change made)

. replace myvar = subinstr(myvar, "VC", "75", .) 
(1 real change made)

. replace myvar = subinstr(myvar, "PC", "35", .) 
(1 real change made)

. list 

    +---------+
    |   myvar |
    |---------|
 1. | 18547A1 |
 2. |  75549F |
 3. |  355297 |
    +---------+

N.B. previous thread How do I find and replace a part of a string variable in Stata?

EDIT The implication of the question appears to be that the strings to be replaced occur just once. If that were ever false, and you only wanted the first occurrence to be replaced, then the advice is to do something like subinstr(myvar, "CC", "18", 1) where the last argument 1 stipulates the (maximum) number of replacements to be made. In other problems this detail could be crucial.

Share:
28,002
Julia
Author by

Julia

Updated on July 09, 2022

Comments

  • Julia
    Julia almost 2 years

    I have observations which list criminal codes as string variables, but not in the format I need. Using Stata 12, I want to replace some substrings in a string variable. For example, I need to change all instances of CC to 18, VC to 75, and PC to 35. Like so:

    Orginal Variable
    CC547A1 | VC549F| PC5297
    
    New Variable
    18547A1 | 75549F | 355297
    

    The characters I need to change are always in the beginning. Some original variables do not need to be changed.

    I tried figuring this out using the substring command, but I just couldn't adapt the code correctly.