Proc transpose

[stancemcgraw]

Hi all,

   I am trying to calculate the time from the last tube feed (tube_stopped) to the time of surgery. Each patient has multiple tube stop times and also multiple surgeries (date_procedure). I want to create a new variable that calculates this called "Time from last tube feed to procedure" in hours....

Data have:

ID                    tube_stopped                date_procedure

7                      12Jul17:02:00:00           13JUL17:13:00:00

7                      01AUG17:15:00:00         01AUG17:22:00:00

7                     20JUL17:07:35:00            20JUL17:22:00:00

1                      29JUL16:01:00:00            29JUL16:23:00:00

1                     12AUG16:16:00:00           12AUG16:22:00:00

Data want:

ID       tube_stopped                 date_procedure               time from last tube feed to procedure

7        01AUG17:15:00:00         01AUG17:22:00:00          2.4

1        12AUG16:16:00:00         12AUG16:22:00:00          5.7

[Reeza]

Use LAG() and basic math, the difference is returned in seconds. To get hours divide by 60 twice.
data want;
set have;
by ID;
Time_Diff = (lag(date_procedure) - tube_stopped)/60/60;
if first.id then call missing(Time_diff);
run;

[Reeza]

Actually upon re-reading this, I may have over complicated this, you can just take the difference of each line and don't need the LAG() function.

[stancemcgraw]

What does the LAG function do?

[Reeza]

LAG gets the previous X row values, I was thinking you need to be looking at the value in the previous row. Lag4 refers to 4 rows back and Lag refers to the previous value