I have trouble efficiently transforming a wrongly coded height variable in my sample.
The variable is coded as follows:
Height
ID1 601
ID1 601
ID1 601
ID3 409
ID3 410
ID4 511
. .
. .
. .
ID100 400
As you can see, the variable is coded as feet and inch simultaneously where 601 equals 6 foot 1 inch, 511 equals 5 ft 11 inches etc.
My goal is to transform these numbers into inches:
replace Height = 48 if Height == 400
replace Height = 49 if Height == 401
replace Height = 50 if Height == 402
replace Height = 51 if Height == 403
.
.
.
replace Height = 83 if Height == 611
How do I efficiently code this with a loop?
The code below assumes that each observation of Height contains exactly three digits:
tostring Height, generate(Height_string)
/* Generate a new variable which is a string-version of Height
(so that we can get the individual digits) */
generate feet = substr(Height_string, 1, 1)
/* From the first character in the string, select one character */
generate inch = substr(Height_string, 2, 2)
/* From the second character in the string, select two characters.
An equivalent alternative would have been
generate inch = substr(Height_string, -2, 2)
which from the second to last character selects two characters */
destring feet inch, replace
/* Convert these two new variables to numeric */
generate tot_inch = feet * 12 + inch
/* Generate a new variable which measure only in inches. */