Expanding to Daily Longitudnal Dataset using two dates

[vldicker]

I need to take my dataset which provides two days and create a line for every date between the two dates for each person.

Example Starting Data

ID            Date1        Date2

1              4/4/16       4/6/16

2              7/8/16       7/9/16

Example Finished Dataset

ID           Date         Date1      Date2

1            4/4/16        4/4/16     4/6/16

1            4/5/16        4/4/16     4/6/16

1            4/6/16        4/4/16     4/6/16

2            7/8/16        7/8/16     7/9/16

2            7/9/16        7/8/16     7/9/16

Date being the new variable in the dataset. 

[FreelanceReinh]

Try this:

data have;
input ID (Date1 Date2) (:mmddyy.);
cards;
1 4/4/16 4/6/16
2 7/8/16 7/9/16
;

data want;
set have;
do Date=date1 to date2;
  output;
end;
format date: mmddyy8.;
run;

Edit: If you don't like the column order in dataset WANT, you can insert the following line after the data want; statement:

retain ID Date Date1 Date2;

[vldicker]

I actually figured it out thanks to Freelance Reinhard

Date Want;

set have;

by id;

if first.id then do;

date=date1;

output;

end;

else;

date=date1;

do unitl (date2);

date+1;

output;

end;

format date mmddyy10.;

run;

[FreelanceReinh]

@vldicker: I'm afraid that your code doesn't work. Typos ("Date" in line 1, "unitl") apart, the UNTIL condition is incorrect: It is equivalent to "date2 is non-missing and not zero." In most cases (incl. your sample data!) this will be satisfied immediately, so that the body of the DO UNTIL loop is executed only once -- irrespective of the difference between date1 and date2.

Edit: Also, the empty ELSE statement does probably not do what you intended (but simply nothing).

Edit 2: Here's a corrected version of your approach:

data want;
set have;
date=date1;
do until(date>date2);
  output;
  date+1;
end;
format date mmddyy10.;
run;

[vldicker]

So your code worked the fastest when I tried what you replied with. But it
makes a double copy of everything (which can be deleted with a proc sort)
but is there a reason it does that?



##- Please type your reply above this line. Simple formatting, no
attachments. -##

[Reeza]

I don't get a double copy.

Post your actual code.

[FreelanceReinh]

Thanks, @Reeza, for stepping in. I must have overlooked the notification that vldicker replied again. There's so much going on in the forum currently.

@vldicker: Sorry for the delayed reply. Are you using two OUTPUT statements (as in your suggested code)? This might explain the "double copies."

[ballardw]

---

@vldicker wrote:
So your code worked the fastest when I tried what you replied with. But it
makes a double copy of everything (which can be deleted with a proc sort)
but is there a reason it does that?

---

Your code had two output statements. So you get one with the first value of id and then because your ELSE doesn't include any statements you get a second for the first value of id.

Date Want;

set have;

by id;

if first.id then do;

date=date1;

output;

end;

else;

date=date1;

do unitl (date2);

date+1;

output;

end;

format date mmddyy10.;

run;