Converting a simple string time variable into minutes

[Kristofer Arestedt]

Hi
I am a beginner at date and time variables in Stata. I understand the principles of date and time variables in Stata but have run into a problem to convert a really simple time variable to minutes. The time variable is a string variable representing the difference between two time points (time of treatment - time of symptom) and include only information about hours and minutes like this example below:
time_dif
00:43
07:28
13:24
04:45

I have read a lot of good instructions to handle time and date variables in Stata but not for a situation with a simple time variable representing a time difference between two time points when no information is available regarding the two different time points (and the date). I suspect that there are a simple solution for this problem but I would really appreciate if someone could give me some advice.

[Nick Cox]

Here as quite often a direct approach using what you knew before age 11 will work fine without record to Stata's date-time functions.

Indeed, if values such as "25:00" are possible this approach is better.

Here are two ways to do it:

clear
input str5 time_dif
"00:43"
"07:28"
"13:24"
"04:45"
end 

gen wanted1 = 60 * real(substr(time_dif, 1, 2)) + real(substr(time_dif, -2, 2))

split time_dif, parse(:) destring 

gen wanted2 = 60 * time_dif1 + time_dif2 

list 

     +----------------------------------------------------+
     | time_dif   wanted1   time_d~1   time_d~2   wanted2 |
     |----------------------------------------------------|
  1. |    00:43        43          0         43        43 |
  2. |    07:28       448          7         28       448 |
  3. |    13:24       804         13         24       804 |
  4. |    04:45       285          4         45       285 |
     +----------------------------------------------------+

. 
.