String - Abbreviating first letter of first name, but leaving surname intact

[Ashwin Sivakumar]

Hi all. How does one go about transforming the string variable name to name2 as seen below:

name	name2
Stephen Curry	S. Curry
Kevin Durant	K. Durant
J.J. Reddick	J.J. Reddick

name is mostly Firstname Lastname as seen, however instances of double initials a la J.J. Reddick and D.J. Augustin do pop up. When looking I have come across ustr and usub but have not been able to get these to work.

Many thanks to all who respond.

[Hemanshu Kumar]

This code works for the example you gave. I don't know what all complications exist in your overall data, so this may or may not suffice:

split name
replace name1 = substr(name1,1,1) + "." if !strpos(name1,".")
gen name_wanted = name1 + " " + name2
drop name?

[Ashwin Sivakumar]

Thank you. split is a command I will definitely be remembering. I was struggling with separating the first and last names.

Also, what does the ! in !strpos do to the command?

[Hemanshu Kumar]

The ! is the NOT operator. If there is a period in name1, strpos will return a positive value, which in logical terms, evaluates to TRUE, and the ! changes that to FALSE (and vice-versa if there is no period in name1).

[Ashwin Sivakumar]

So it does the counter of the original command. Many thanks. Have a nice day.

[William Lisowski]

For Stata users comfortable with regular expression matching syntax, this can be done using a single match-and-replace.

. generate name2 = ustrregexrf(name, "^([A-Z])[^ \.]* ", "$1. ")

. list, clean

                name          name2  
  1.   Stephen Curry       S. Curry  
  2.    Kevin Durant      K. Durant  
  3.    J.J. Reddick   J.J. Reddick  
  4.     W. Lisowski    W. Lisowski