BookmarkSubscribeRSS Feed
knveraraju91
Barite | Level 11

Dear,

 

I got output i need but the pgm looks lengthy. Please suggest some alternative way or sql code. Thank you

 

I need to select for day variable value and create impday  for each id by selecting  first obs after last obs where flag='N'.

for id=1,  the first obs after last obs where flag='N' is   day=5. so IMPDAY=5

for id=2, the first obs after last obs where flag='N' is   day=5. so IMPDAY=5

for id=3, there is no obs where flag='N' so i need to take first obs

for id=4  there is one record so i select

output needed

id flag day impday

1 Y 1 5
1 Y 2 5
1 Y 3 5
1 N 4 5
1 Y 5 5
1 Y 6 5
2 N 1 5
2 Y 2 5
2 Y 3 5
2 N 4 5
2 Y 5 5
2 Y 6 5
3 Y 1 1
3 Y 2 1
3 Y 3 1
4 Y 1 1

data one;
input id flag$ day;
datalines;
1 Y 1
1 Y 2
1 Y 3
1 N 4
1 Y 5
1 Y 6
2 N 1
2 Y 2
2 Y 3
2 N 4
2 Y 5
2 Y 6
3 Y 1
3 Y 2
3 Y 3
4 Y 1
;

proc sort data=one;
by id descending flag;
run;

data two;
set one;
by id descending flag;
if last.id;
p_flag=flag;
p_day=day;
drop flag day;
run;

data three;
merge one(in=a) two(In=b);
by id;
if a;
if p_flag='N' and day lt p_day then delete;
run;

proc sort data=three;
by id flag;
run;

data four;
set three;
by id flag;
if flag='N' and first.id ^=last.id then delete;
run;

proc sort data=four;
by id flag;
run;
data five;
set four;
by id flag;
if first.id;
impday=day;
keep id impday;
run;

data six;
merge one(in=a) five;
by id;
if a;
run;
6 REPLIES 6
ballardw
Super User

Since Proc SQL does not really have a good sense of "order" of records your "first obs after Last obs flag='N' " it is unlikely to be a good tool alone.

 

What do you want when the last observation for an ID has flag="N"? The next observation is from a different ID.

knveraraju91
Barite | Level 11
Thank you very much for support. i set to missing if id has only one record and has "N'
Reeza
Super User

Personally I prefer a data step but here are two approaches.

 

proc sort data=one; 
by id day;
run;

data temp;
set one (where=(flag='N'));
by ID;
impDay = day+1;
if last.ID;
keep id impday;
run;

data want;
merge one temp;
by id;
if missing(impday) then impday = 1;
run;

SQL option


proc sql;
create table want_sql as
select t1.*, coalesce(t2.max_day, 1) as impday
from one as t1
left join (select id, max(day)+1 as max_day from one where flag='N' group by ID) t2
on t1.id=t2.id;
quit;
knveraraju91
Barite | Level 11

Thank you very much for support. The code is working for some OBS only if DAY values are sequential order. But if day variable values are not sequential order impday= DAY+1 is not giving me correct output as mentioned in code. I need to take next OBS after last obs having flag='N'. Please suggest.

Reeza
Super User
Please provide sample data that reflects your business requirements.
ballardw
Super User

@knveraraju91 wrote:

Thank you very much for support. The code is working for some OBS only if DAY values are sequential order. But if day variable values are not sequential order impday= DAY+1 is not giving me correct output as mentioned in code. I need to take next OBS after last obs having flag='N'. Please suggest.


Since all of your example data shows DAY as sequential you implied that was the case.

If you do not provide representative data it is hard to generate a general solution.

 

And per my previous question what if the last value when you have more than one value in an id group that the flag is 'N'. You only answered for a specific case of exactly one value.

SAS Innovate 2025: Call for Content

Are you ready for the spotlight? We're accepting content ideas for SAS Innovate 2025 to be held May 6-9 in Orlando, FL. The call is open until September 25. Read more here about why you should contribute and what is in it for you!

Submit your idea!

How to Concatenate Values

Learn how use the CAT functions in SAS to join values from multiple variables into a single value.

Find more tutorials on the SAS Users YouTube channel.

Click image to register for webinarClick image to register for webinar

Classroom Training Available!

Select SAS Training centers are offering in-person courses. View upcoming courses for:

View all other training opportunities.

Discussion stats
  • 6 replies
  • 805 views
  • 1 like
  • 3 in conversation