Same Observation

[Smitha9]

Hi,

I have a dataset

ID  Name         DOB

1   Sarah    07/29/1970

1       .                .

1   Sarah      07/29/1970

I want to fill the missing values and dataset which I have has 10000's ID's.

ID  Name         DOB

1   Sarah    07/29/1970

1   Sarah     07/29/1970               

1   Sarah      07/29/1970

can I do this in SAS?

thanks in advance.

[PaigeMiller]

From now on, @Smitha9 please provide data as working SAS data step code, as I show here. Please test your code to confirm that it actually works. What you have provided is not SAS data step code.

data have;
input ID  Name $8. DOB :anydtdte.;
infile cards missover;
cards;
1   Sarah    07/29/1970
1                
1   Sarah    07/29/1970
;

Here is my solution

data want;
    set have;
    retain name1 dob1;
    if not missing(name) then name1=name;
    if not missing(dob) then dob1=dob;
    drop name dob;
    format dob1 mmddyy10.;
run;

[Kurt_Bremser]

It seems that one also needs to take ID into account:

data have;
input ID Name :$8. DOB :mmddyy10.;
format dob yymmdd10.;
infile cards missover;
cards;
1 Sarah 07/29/1970
1  .
1 Sarah 07/29/1970
;

data want;
set have;
by id;
retain _name _dob;
if first.id or not missing(name)
then _name = name;
else name = _name;
if first.id or not missing(dob)
then _dob = dob;
else dob = _dob;
drop _dob _name;
run;

[Astounding]

Here's a different approach.  It has a major advantage in that it fills in missings for all your variables:

data want;
   update have (obs=0) have;
   by id;
   output;
run;

So the program remains short no matter how many variables you have.  The disadvantage:  you need a sorted data set for this to work.  Another consideration:  it requires extra work if you want to fill in missings for only some of your variables but not for others.