Question

I have the following data set structure:

date    time_in_hours   price   
Sep 03 08   9.76            98  
Sep 03 08   10.43           97  
Sep 03 08   10.98           96      
Sep 03 08   11.48           99      
Sep 04 08   2.35            98      
Sep 04 08   2.58            98.45       
Sep 04 08   3.45            96.3        
Sep 04 08   3.89            96.25       
Sep 04 08   4.18            100     
Sep 05 08   12.65           101     
Sep 05 08   12.96           100.25      
Sep 05 08   13.25           104.35      
Sep 05 08   13.78           98      

My data is for the years 2008 and 2009. It contains a total of 504 trading days.
My objective is to interpolate prices at every half hour (eg 9.5 10 10.5 11 11.5…etc.) only for time interval between 9.5 and 16.

I have been struggling with the command interpolate / aggregate given that I must interpolate for a specific time interval for each calendar date. My final output must also contain the date, time, and price. Something like this:

date    time_in_hours   price   
Sep 03 08   10           98  
Sep 03 08   10.5         97  
Sep 03 08   11           96      
Sep 03 08   11.5         99      
Sep 04 08   2.5          98      
Sep 04 08   3            98.45     

Answer 1

The code below gives you the output you have stated, but based on all comments above I am not sure it is going to solve your problem completely. Note that round(x, 0.5) takes 0.25 as the boundary – so 2.74 becomes 2.5 while 2.75 becomes 3.

data test;
infile datalines dsd;
input date :$20. time_in_hours price;
datalines;
Sep 03 08,9.76,98
Sep 03 08,10.43,97
Sep 0308,10.98,96
Sep 03 08,11.48,99
Sep 04 08,2.35,98
Sep 04 08,2.58,98.45
Sep 04 08,3.45,96.3
Sep 04 08,3.89,96.25
Sep 04 08,4.18,100
Sep 05 08,12.65,101
Sep 05 08,12.96,100.25
Sep 05 08,13.25,104.35
Sep 05 08,13.78,98
;
run;

proc print;
run;

data test2;
    set test(rename = (time_in_hours = old_time_in_hours));
    time_in_hours = round(old_time_in_hours, 0.5);
    if (9.5 <= time_in_hours <= 16);
run;

proc print;
run;