identifying incomplete observations

[jkf91]

Hi all,

I have a specific problem which I'm not sure how to solve. Suppose I have the following data called all.all:

I want to identify id's such as id 3. Eventually, I need to create a data subset that includes id's that:

1. do not have a change in household_size over years (e.g., id 2 had household_size change between 2004 and 2005.)

2. do nto have a gap year (e.g., id 1 is missing 2005)

Thanks,

Chris

[art297]

Here is one way:

proc sql noprint;

  create table want1 as

    select *

      from have (where=(year in (2004, 2005)))

        group by id

          having min(household_size) eq

                 max(household_size) and

                 count(*) eq 2

  ;

  create table want2 as

    select *

      from have

        group by id

          having max(year)-min(year)+1 eq

                 count(*)

  ;

quit;

[Tom]

data want ;

    yeargap = 0;

    sizechange=0;

   do until (last.id) ;

       set have ;

       by id ;

      yeargap = max(yeargap , year ne lag(year) and not first.id);

       sizechange = max(sizechange, size ne lag(size) and not first.id);

    end;

   keep id yeargap sizechange;

run;

[Ksharp]

How about:

data temp;
input id year household_size;
cards;
1 2004 2
1 2006 2
1 2007 2
1 2008 2
2 2003 3
2 2004 3
2 2005 2
3 2003 3
3 2004 3
3 2005 3
;
run;
proc sql noprint;
 create table want as
  select *
   from temp
    group by id
     having count(distinct household_size) eq 1 and range(year) eq count(year)-1;
quit;

Ksharp