How to create dynamic suffix on a numbered variable list?

[gsouther]

Hello, I am just learning SAS and so far have not found a way to dynamically create variables based on a count of a character in a string. I am parsing out a string from Active Directory (for example: CN=N123456,OU=Users,OU=Admins,DC=D01,DC=com). I have successfully created collumns for each item following a comma, but it is not dynamic. I have been running a proc sql query to find the Active Directory string with the most commas, and then creating an array based on that number.

In the file I am currently working with, the largest number of commas is 7, so I need 8 columns. I have been creating varialbes like this:

array parsed_vars(8)$50 VAR1-VAR8;

I want to store the number of collumns I need based on my comma count and pass that into the array, something like this:

array parsed_vars(nmbrvar)$50 VAR1-VAR||nmbrvar;

The log shows this:

ERROR: Missing numeric suffix on a numbered variable list (OU1-OU).

Is there a way to do this? I am working in SAS Enterprise Guide 5.1.

Thanks.

[Kurt_Bremser]

In your proc sql, use the into: clause to store your result into a macro variable.

Like

proc sql noprint;
select max(count) into : maxcount from ....;
quit;

You can then use maxcount in further coding:

data something;
set have;
array vars {&maxcount} var1-var&maxcount;
.....
run;

[gsouther]

Thanks for the response! I was missing the into: clause. It seems to be getting a little further, but I am still getting the error about missing the numeric suffix.

Code:

proc sql;
select max(count(distinguishedname,","))+1 into : commacount from have;
quit;


data something;
set have;
array parsed_vars{&commacount}$50 VAR1-VAR&commacount;
run;

28         data something;
29         set have;
30         array parsed_vars{&commacount}$50 VAR1-VAR&commacount;
NOTE: Line generated by the macro variable "COMMACOUNT".
30         VAR       9
                    _
                    22
                    200
ERROR: Missing numeric suffix on a numbered variable list (VAR1-VAR).
ERROR: Too few variables defined for the dimension(s) specified for the array parsed_vars.
ERROR 22-322: Syntax error, expecting one of the following: a name, (, ;, _ALL_, _CHARACTER_, _CHAR_, _NUMERIC_.  

ERROR 200-322: The symbol is not recognized and will be ignored.

[Astounding]

You're most of the way there.

The problem is that SQL must perform a numeric to character conversion, to feed the numeric value into &COMMACOUNT.  This conversion generates leading blanks in the value of &COMMACOUNT.  One simple way to remove those leading blanks would be:

%lead commacount = &commacount;

That would go between the SQL and the DATA step, and should take care of the problem.

[Kurt_Bremser]

Some autocompletion seems to have played a trick on you.

%lead commacount = &commacount;

should be

%let commacount = &commacount;

[Astounding]

That's just my brain trying to function before the first cup of coffee.  Absolutely, use %LET not %LEAD.

[gsouther · Accepted solution]

Thanks for the help! I have it working now with the code below.

proc sql;
select max(count(distinguishedname,","))+1 into : commacount from have;
quit;

%let commacount = &commacount;

data something;
set have;
array parsed_vars{&commacount}$50 VAR1-VAR&commacount;
run;

[RW9]

Personally I would normalise your data, i.e. put each data item into a separate observation.  If necessary you could then transpose them up or work with it as is.

data temp;
  str="CN=N123456,OU=Users,OU=Admins,DC=D01,DC=com";
  do i=1 to countw(str,",");
    word=scan(str,i,",");
    output;
  end;
run;
proc transpose data=temp out=t_temp prefix=word;
  var word;
  id i;
  idlabel i;
run;

[rogerjdeangelis]

HAVE

Up to 40 obs WORK.HAVE total obs=10

Obs    CMADLM

  1    aaaa,aaaa,aaaa,aaaa,1stwrd
  2    aaaa,aaaa,aaaa,aaaa,aaaa,aaaa,aaaa,aaaa,aaaa
  3    aaaa,aaaa,aaaa
  4    aaaa,aaaa,aaaa,aaaa,aaaa,aaaa,aaaa,aaaa,aaaa
  5    aaaa,aaaa
  6    aaaa,aaaa,aaaa,aaaa
  7    aaaa,aaaa
  8    aaaa,aaaa,aaaa,aaaa,aaaa,aaaa
  9    aaaa,aaaa,aaaa,aaaa,aaaa,aaaa,aaaa
 10    aaaa,aaaa,aaaa,aaaa,aaaa,aaaa,aaaa,aaaa,aaaa

WANT


Up to 40 obs WORK.TEMP total obs=10

Obs     V1      V2      V3      V4       V5       V6      V7      V8      V9

  1    aaaa    aaaa    aaaa    aaaa    1stwrd
  2    aaaa    aaaa    aaaa    aaaa    aaaa      aaaa    aaaa    aaaa    aaaa
  3    aaaa    aaaa    aaaa
  4    aaaa    aaaa    aaaa    aaaa    aaaa      aaaa    aaaa    aaaa    aaaa
  5    aaaa    aaaa
  6    aaaa    aaaa    aaaa    aaaa
  7    aaaa    aaaa
  8    aaaa    aaaa    aaaa    aaaa    aaaa      aaaa
  9    aaaa    aaaa    aaaa    aaaa    aaaa      aaaa    aaaa
 10    aaaa    aaaa    aaaa    aaaa    aaaa      aaaa    aaaa    aaaa    aaaa


* The maximum size the array can be is 656 because 656*50=32800;
* 50 was mentioned in the instructions - but could set higher 4096;
* without performance penaly?;

data interim;
  retain cmacnt 0;
  set have(in=have);
  array vars[*] $50 v1-v656;
  do cma=1 to (countc(cmadlm,',')+1);
     vars[cma]=scan(cmadlm,cma,',');
  end;
  if cmacnt < cma-1 then cmacnt=cma-1;
  call symputx('cmacnt',put(cmacnt,4.));
run;quit;

data temp;
  set interim(keep=v1-v&cmacnt);
run;quit;

[rogerjdeangelis]

Code to create have

data have;
 length cmadlm $500;
 retain cmadlm '1stwrd';
 do str=1 to 10;
    too=10*uniform(57371);
    do cma=1 to too;
      cmadlm=catx(',','aaaa',cmadlm);
    end;
    output;
    cmadlm=' ';
 end;
 keep cmadlm;
run;quit;

[rogerjdeangelis]

Correction I don't know what I was thinking but 656 is not large enough, it assumes the string has at least 50 bytes between commas.

Oh well, I thinl 16384 is the max.