Determining increase and decrease values

[Dregerator]

Hi I have this table and I want to determine when dose increases and dose decreases programatically 

GROUP	DOSE	START	END
1	40	8/24/2022	9/22/2022
1	70	9/23/2022	10/7/2022
1	70	10/9/2022	10/13/2022
1	64	10/15/2022	10/19/2022
1	48	10/21/2022
2	35	8/1/2022	8/14/2022
2	70	8/15/2022	8/19/2022
2	80	8/20/2022	8/30/2022
2	90	8/31/2022	9/11/2022
2	90	9/12/2022	9/25/2022
2	50	9/26/2022	10/2/2022
2	35	10/4/2022	10/10/2022
2	35	10/11/2022	10/25/2022
2	20	10/26/2022	11/1/2022
2	10	11/2/2022

What I want the final product to look like is this 

GROUP	DOSE	START	END	Start	End	Increase	Decrease
1	40	8/24/2022	9/22/2022	Y
1	70	9/23/2022	10/7/2022			Y
1	70	10/9/2022	10/13/2022
1	64	10/15/2022	10/19/2022				Y
1	48	10/21/2022					Y
2	35	8/1/2022	8/14/2022	Y
2	70	8/15/2022	8/19/2022			Y
2	80	8/20/2022	8/30/2022			Y
2	90	8/31/2022	9/11/2022			Y
2	90	9/12/2022	9/25/2022
2	50	9/26/2022	10/2/2022				Y
2	35	10/4/2022	10/10/2022				Y
2	35	10/11/2022	10/25/2022				Y
2	20	10/26/2022	11/1/2022				Y
2	10	11/2/2022					Y

Any help is greatly appreciated 

[PeterClemmensen]

Why is End never equal to 'Y' ?

[Dregerator]

There is no end date for the last record for group 1 and group 2 . In the future, there will be a date but at the moment it is blank

[PeterClemmensen · Accepted solution]

@Dregerator ok. Then simply take out the line setting the _end variable. 

Let me know if my code works for you.

[PeterClemmensen]

This should get you going.

data have;
input GROUP DOSE (START END)(:mmddyy10.);
infile datalines missover;
format START END mmddyy10.;
datalines;
1 40 8/24/2022   9/22/2022   
1 70 9/23/2022   10/7/2022   
1 70 10/9/2022   10/13/2022  
1 64 10/15/2022  10/19/2022  
1 48 10/21/2022              
2 35 8/1/2022    8/14/2022   
2 70 8/15/2022   8/19/2022   
2 80 8/20/2022   8/30/2022   
2 90 8/31/2022   9/11/2022   
2 90 9/12/2022   9/25/2022   
2 50 9/26/2022   10/2/2022   
2 35 10/4/2022   10/10/2022  
2 35 10/11/2022  10/25/2022  
2 20 10/26/2022  11/1/2022   
2 10 11/2/2022               
;

data want;
   set have;
   by group;
   
   if first.group then _start = 'Y';
   if last.group  then _end   = 'Y';

   increase = ifc(dif(dose) > 0 & first.group = 0, 'Y', '');
   decrease = ifc(dif(dose) < 0 & first.group = 0, 'Y', '');

run;

[Moonlight_26]

Woww, I have a case similar with this and I can solve my problem by using your scripting model. Thank you very much.