Getting past weeks data

[Vasundha]

Right

[PeterClemmensen]

If so, then try this

Data have;
Input in_date : date9. readings;
format in_date date9.;
Datalines;
04Nov2022 40
08Nov2022 89
09NOV2022 61
10NOV2022 60
11NOV2022 67
15NOV2022 45
16NOV2022 15
17NOV2022 12
18NOV2022 10
;
run;

data want;
  do p = max(1, n - 6) to n;
    set have nobs = n point = p;
    output;
  end;
  stop;
run;

[Vasundha]

I really appreciate your efforts however I was wondering that can't it be possible in input table in a data step?.

[Kurt_Bremser]

To evaluate data with SAS, it must first be brought into SAS. First import, then use one of the methods suggested, then subset.

[Vasundha]

Alright, Thanks!.

[PeterClemmensen]

If by 'input', you mean directly in the set(someoption = )statment, then no.

[Vasundha]

Got it, Thank you for your precious time!.

[Kurt_Bremser]

Brute force approach:

proc sort
  data=have (keep=date)
  out=dates
  nodupkey
;
by descending date;
run;

data _null_;
set dates point=7;
call symputx("cutoff",date);
stop;
run;

data want;
set have;
where date ge &cutoff.;
run;

But the PROC RANK method is more efficient.