Agreed, @Ksharp 's reply is probably the best reply so far to match the specs. ... View more

Aligned Box Criterion is available in the HP Cluster node in SAS Enterprise Miner. It will determine the optimum number of clusters. Here's a video that talks about using this option, along with using CCC and gap methods https://www.youtube.com/watch?v=NZpNTkfT47c ... View more

I only see a dataset with one variable and one observation. Since this is no good on its own, I was asking for the bigger picture. But maybe you don't have one.   To suggest an alternative (as you wanted in your first post), we need to know what you're aiming at. SAS will not resolve anything (no macro references, no function call, nothing) in a datalines block. And your SQL approach works, as demonstrated. But we have no idea why it would cause such a big change to existing code, as we don't see anything of that. ... View more

  Here is a macro that will add the keyword WHERE to the beginning if the value does not already start with the word WHERE. %macro chk(where); %if %length(&where) %then %do; %if "WHERE " ne "%qsubstr(%qupcase(&where) 12345,1,6)" %then %let where=where &where ; %end; %put &=where ; %mend; %chk(); %chk(fred); %chk(wherefred); %chk(where age>10); The normal ISBLANK macro I have seen tests if the value is empty or only contains spaces. I normally just use %LENGTH() to test if the string is empty.  To see the difference try calling the test macro with a macro quoted blanks.   %chk(%str( )); Note that you can defined the parameter as positional so that you can call the macro without listing the parameter name.  But you can still call it using the keyword if you want.  You just can't call by position when the parameter is defined as keyword. %chk(where=where age>10); %chk(where=age>10); Example results: 88 %chk(); WHERE= 89 %chk(fred); WHERE=where fred 90 %chk(wherefred); WHERE=where wherefred 91 %chk(where age>10); WHERE=where age>10 92 %chk(%str( )); WHERE=where 93 %chk(where=where age>10); WHERE=where age>10 94 %chk(where=age>10); WHERE=where age>10       ... View more

Coding is much easier in data steps.  Plus the DO loop is much more powerful than the %DO loop. Not sure what you want to do but I can generate your pattern of K and ALPHA values using data steps like this.   I added extra DO loops so you can play with settings of two input numbers and see how it works. data _null_; do sixpack=6; do char_var_cnt=12,24,17 ; put / sixpack= char_var_cnt= / 12*'-' ; alpha=1 ; do k=1 by 1 while (alpha <= char_var_cnt) ; put k= @; do i=1 to ceil(char_var_cnt/sixpack) while (alpha <= char_var_cnt); put alpha= @; alpha+1; end; put; end; end; end; run; sixpack=6 char_var_cnt=12 ------------ k=1 alpha=1 alpha=2 k=2 alpha=3 alpha=4 k=3 alpha=5 alpha=6 k=4 alpha=7 alpha=8 k=5 alpha=9 alpha=10 k=6 alpha=11 alpha=12 sixpack=6 char_var_cnt=24 ------------ k=1 alpha=1 alpha=2 alpha=3 alpha=4 k=2 alpha=5 alpha=6 alpha=7 alpha=8 k=3 alpha=9 alpha=10 alpha=11 alpha=12 k=4 alpha=13 alpha=14 alpha=15 alpha=16 k=5 alpha=17 alpha=18 alpha=19 alpha=20 k=6 alpha=21 alpha=22 alpha=23 alpha=24 sixpack=6 char_var_cnt=17 ------------ k=1 alpha=1 alpha=2 alpha=3 k=2 alpha=4 alpha=5 alpha=6 k=3 alpha=7 alpha=8 alpha=9 k=4 alpha=10 alpha=11 alpha=12 k=5 alpha=13 alpha=14 alpha=15 k=6 alpha=16 alpha=17 ... View more