Here's another solutions:   data have; input ID    '200101'n '200102'n  '200103'n; cards; 1 1 . 1 2 . . 1 3 . 1 1 ; proc transpose data=have out=tran(rename=_NAME_ = DATE);by id; proc sort data=tran;by id descending col1 date; data prep; set tran; by id descending col1 date; if first.id then first = date; run; proc sql; create table want as select distinct a.*, b.First from have a left join      prep b on a.id = b.id where not missing(b.first) order by id; ... View more

Not necessarily the transpose, I was only suggesting that the transpose 'might' get you unintended results.  I assume you have to do the sort.  Think about what the lag function does, then think about a sort.  Judging by your picture you want to sort before using the lag function. ... View more

Depending on what you want to see you should use the lag function last, I see the proc sort has already been addressed but you also have a transpose after the lag which might get you unintended results. ... View more

or this: data want(drop=flag); do until(last.product); set have; by product; if date = '30sep2015'd and mtd_sales > goal then flag = 1; end; do until(last.product); set have; by product; if flag = 1 then Color = 'GREEN'; else color = 'RED'; output; end; run; ... View more

If you are looking for the code and not the report here you go:   data have; infile cards; input product$ date mtd_sales goal; informat date date9.; format date date9.; cards; Handbags 21-Sep-15 23 50 Handbags 22-Sep-15 25 50 Handbags 23-Sep-15 28 50 Handbags 24-Sep-15 29 50 Handbags 25-Sep-15 33 50 Handbags 26-Sep-15 34 50 Handbags 27-Sep-15 37 50 Handbags 28-Sep-15 40 50 Handbags 29-Sep-15 44 50 Handbags 30-Sep-15 45 50 Shoes    21-Sep-15 38 70 Shoes    22-Sep-15 44 70 Shoes    23-Sep-15 46 70 Shoes    24-Sep-15 47 70 Shoes    25-Sep-15 49 70 Shoes    26-Sep-15 57 70 Shoes    27-Sep-15 63 70 Shoes    28-Sep-15 68 70 Shoes    29-Sep-15 72 70 Shoes    30-Sep-15 74 70 ; proc sql; create table want as select product,date,mtd_sales,goal,max(color) as color from ( select *, case when date = '30sep2015'd and mtd_sales > goal then 'GREEN' when date = '30sep2015'd and mtd_sales < goal then 'RED' end as Color from have) group by product; ... View more

Does this get you what you are looking for?   proc sql noprint; select distinct memname into :namelist separated by ' ' from dictionary.columns where libname = 'WORK' and memname ne '_PRODSAVAIL'; %put &namelist; ... View more

If I understand your question correctly then yes.  Put a comma in front and behind the case statement as if it were any other item you are selecting from the dataset. ... View more

For if then look up pl/sql decode() or case statements. ... View more