Creating a sample with thousands of criteria.

[Allan]

Hi,

I have to create a sample DATASET where there are 'n' records for every day between 20160101 and 20210630 about 2008 days.
Now there are various ways I've thought of doing this.
I could write a MACRO to generate a CHOOSESETS that has a criteria for every day
e.g.

Code: Select all

x := CHOOESETS(inputDS,date = '20160101' => n
                      ,date = '20160102' => n
etc


I could SORT by date then ITERATE retaining a counter and just SKIP when inputs counter passes 'n'.
But one feels these are all to complicated and I'm missing a very simple way to do this.

By the way I can error if there are no records for a day by LEFT ONLY join on a datasets of complete date range, so that's not an issue.
Any Ideas?

Allan

[Allan]

AH Of Course

Just a inner JOIN against a dataset with list of dates with KEEP(n) as a qualifier.

[rtaylor]

Allan,

You might also do it this way:

Code: Select all

IMPORT Std;
Std.Date.Date_t StartDate := 20210101;
Std.Date.Date_t EndDate   := 20210110;
n := 10;

nDays   := Std.Date.DaysBetween(StartDate,EndDate) + 1;
NumRecs := nDays * n;
Std.Date.Days_t StartDateJ := Std.Date.FromGregorianDate(StartDate);

OutRec := RECORD
  Std.Date.Date_t ThisDate;
  UNSIGNED1       WhichRec;
  STRING10        OtherStuff;   
END;
OutRec XF(INTEGER C) := TRANSFORM
  Cmod := C % n;
  Cday := IF(Cmod <> 0,TRUNCATE(C/10) + 1,TRUNCATE(C/10));
  SELF.ThisDate := Std.Date.ToGregorianDate(StartDateJ + (Cday-1));
  SELF.WhichRec := IF(Cmod <> 0,Cmod,n);
  SELF.OtherStuff := 'junk ' + C;
END;
ds := DATASET(NumRecs,XF(COUNTER));
ds;

HTH,

Richard