how to incorporate 'NODUPKEY' in hash?

[sassharp]

need to incorporate 'NODUPKEY' in hash short as in proc sort nodupkey. Any examples?

[sassharp]

http://www.lexjansen.com/pharmasug/2011/tt/pharmasug-2011-tt15.pdf

[sassharp]

http://www2.sas.com/proceedings/sugi30/236-30.pdf

[Haikuo]

data have;

input id  var @@;

cards;

1 2 1 3 10 6 4 5 10 7 8 2

;

data _null_;

  if _n_=1 then do;

dcl hash h( ordered: 'a');

h.definekey('id');

h.definedata('id','var');

h.definedone();

   end;

   do until (done);

   set have end=done;

   _rc=h.add();

      end;

h.output(dataset:'want');

   stop;

   run;

Regards,

Haikuo

[FriedEgg]

The hash object by default drops duplicate keys.  This is controlled by the duplicate option which tells SAS whether to replace or error when a duplicate key is encountered (when loading a dataset into a hash using the dataset option).  To avoid this use the multidata:'y' option in the declare statement.

http://support.sas.com/documentation/cdl/en/lrdict/64316/HTML/default/viewer.htm#a002576871.htm

Check out the post by MikeZdeb in the link for a great list of reference papers for learning about Hash objects in SAS:

http://communities.sas.com/thread/33376?tstart=0

[Haikuo]

Reminded by FriedEgg's comments "The hash object by default drops duplicate keys.", add() method is not needed for this purpose if choose to load the dataset once for all. However, if chose the last record when duplicating, replace() may still be needed:

data have;

input id  var @@;

cards;

1 2 1 3 10 6 4 5 10 7 8 2

;

   data _null_;

  set have (obs=1);

dcl hash h( dataset: 'have', ordered: 'a');

h.definekey('id');

h.definedata('id','var');

h.definedone();

h.output(dataset:'want');

   run;

This one feels better.

Regards,

Haikuo

[FriedEgg]

I may not have made myself clear enough:

keep the first incidence of key

dcl hash h(dataset:'have', ordered:'a');

keep the last incidence of key

dcl hash h(dataset:'have', ordered:'a' ,duplicate:'r');

keep all incidience of key

dcl hash h(dataset:'have', ordered:'a' ,multidata:'y');

fail if duplicates of key exist in loading file

dcl hash h(dataset:'have', ordered:'a' ,duplicate:'e');

[Haikuo]

Learned and Agreed!