Create Macros Conditionally

[sasmaverick]

I am trying to create multiple macro variables based on various conditions. Below is my code. I am not sure whether this is the most efficient way.

*to get total number of obs and max value of count variable where count<5;

proc sql noprint;

select count(*),max(count)

into : count1, count2

from dataset1

where count<5

;quit;

*to get total number of obs and max value of count variable where count>=5;

proc sql noprint;

select count(*),max(count)

into : count1, count2

from dataset1

where count>=5

;quit;

*count of name where count>=5;

proc sql;

select count (distinct name) into:cnt

from dataset1

where count>=5;

quit;

*count of name where count<5;

proc sql;

select count (distinct name) into:cnt

from dataset1

where count<5;

quit;

*get names into macro where count<5;

proc sql

select distinct name into : name1-:name&sysmaxlong

from dataset1

where count<5;

quit;

repeat the above step for count>=5

Is there a shorter/more concise way for the above. Will appreciate inputs on a more efficient approach.

Thanks!all

[Ksharp]

No need for so many proc sql block, only one is enough . and combine these sql statement which are under the same condition.

proc sql;

........

.......

select count(*),max(count),count (distinct name)

into : count1, :count2, :cnt

from dataset1

where count>=5

;

............

............

quit;

[jwillis]

Adding to your code:

%macro choose(type=);

proc sql;

........

.......

select count(*),max(count),count (distinct name)

into : count1&type., :count2&type., :cnt&type.

from dataset1

%if &type = eq %then %do;

where count = 5

%end;

%if &type = lt %then %do;

where count < 5

%end;

%if &type = gt %then %do;

where count > 5

%end;

;

quit;

%mend choose;

%choose(type=eq);

%choose(type=lt);

%choose(type=gt);

[RW9]

Hi,

IMO the answer to:

" I am not sure whether this is the most efficient way."

would be no.  Whenever I need to deal with more than one (or even one if its dates or such like) then I would put parameters into a dataset.  It is then simply a matter of merging on the parameter dataset to use them.  Avoids all that nasty:

&&&%str(&&&av_&&i.) nonsense with simple to understand merging.

Most languages provide some sort of macro facility, but they are used very rarely and only for specific things.  The opposite is true in SAS, most of the code seems to be in macros even though with a bit of thought 90% of it can be done in straight code.