Calculate difference between columns for rows when a string condition is found

[ACLAN]

I'm trying to correct the values in a specific columns for rows in which a specific condition is found.

Using the following dataset as example:

data dttest;

input patient_id $ patient_name $ (enttime draw1 tdiff) (time8. +1);

format enttime draw1 tdiff time8.;

datalines;

18C020 Garcia 11:45:00 11:50:00 00:07:00

17C222 Gibson 11:43:00 11:47:00 00:04:00

18C023 Knapp 11:42:00 11:45:00 00:43:00

10C032 Mueller 09:11:00 11:47:00 02:36:00

;

run;

For the rows where patient_id has "18C", I would like to correct the values in column tdiff by subtracting the time in colum draw1 from the time in column enttime.

I have tried IF/THEN and WHERE to select the rows but I always have an error. Not sure if it is a good idea to create an array.  

Anyone has a simple solution for this?

[Jagadishkatam · Accepted solution]

Please try

data dttest;
input patient_id $ patient_name $ (enttime draw1 tdiff) (time8. +1);
format enttime draw1 tdiff time8.;
datalines;
18C020 Garcia 11:45:00 11:50:00 00:07:00
17C222 Gibson 11:43:00 11:47:00 00:04:00
18C023 Knapp 11:42:00 11:45:00 00:43:00
10C032 Mueller 09:11:00 11:47:00 02:36:00
;
run;

data want;
set dttest;
if index(patient_id,'18C') then tdiff=draw1-enttime;
format tdiff time8.;
run;

[ACLAN]

It works thanks. I forgot to define the format