Does that mean that it takes 50 variables to uniquely identify an observation in your source dataset?  So that you are worried about the space/time to store 3 million * 4 * (length of 50 variables plus the _name_ and COL1 variables) bytes of data? From your original request you do not seem to care about those variables so it should be fine not to include them in the transposed data. You can always use a data step view to construct the transposed data to feed to the PROC SQL code so that you do not need to store that intermediate dataset at all. data test2 / view=test2 ;    length _name_ $32 col1 8 ;    set test ;   array vars f1-f4 ;    do i=1 to dim(vars);      _name_ = vname(vars(i));     col1 = vars(i);      output;    end;   keep _name_ col1; run; Or does it mean that you want to search over 50 variables instead of 4. The difference in space to store 50 variables per observation versus one variable in 50 times as many observations is not much. ... View more

At least with Office 2007 WORD will still display immediately in the file the field result values that you store using the \fldrstl tag. For example here as simple one line RTF file that contains Page 1 of 1 done with commands. With SAS I have no idea how it could know what numbers to store in the file for a normal RTF output.  SAS would need to make a guess at how WORD was going to align the text onto the pages to figure out how many pages there are and what current page this instance of the PAGE field command is on. {\rtf1 {Page } {\field{\*\fldinst { PAGE }}{\fldrslt {1}}} { of } {\field{\*\fldinst { NUMPAGES  }}{\fldrslt {1}}} } ... View more

I still have no idea what you are actually trying to do, but try this method using PROC TRANSPOSE to convert your 4 variables into 4 observations so that we can query over them. proc transpose data=test out=test2 ;   by name add notsorted ;   var f1-f4 ; run; proc sql noprint;   create table test2 as     select distinct col1 as value,_name_,col1     from test2     group by col1     having count(distinct _name_) > 2     order by 1,2   ; quit; proc transpose data=test2 out=want(drop=_name_) ;   by value;   id _name_;   var col1 ; run; proc print; run; output: Obs    value    f1    f2    f3    f4 1       1       1     1     1     1 2       2       2     2     .     2 3       4       4     4     4     4 ... View more

The outer most quotes are the ones that determine if macro references are resolved. Try: %let date1=01JAN2011; %let x1="The date value is '&date1' "; %let x2='The date value is "&date1" '; %put x1=%superq(x1); %put x2=%superq(x2); ... View more

Your data structure does not match the way you are referring to the information. You talk about the values inside individual variables as if they were the observations. If you restructure the data into two variables, one with the value and one with the column label then the query becomes easier. data test; do col='f1','f2','f3','f4';    input value @; output; end; cards; 1 2 3 4 4 5 1 2 8 1 9 0 3 4 4 1 9 12 12 14 2 16 17 18 run; proc sql noprint ;   create table want as     select distinct value     from test     group by value     having count(distinct col) > 2   ; quit; proc print; run; ... View more

I have a simple one that I wrote a couple of years ago based on ideas in a paper by Craig Roberts and John R. Gerlach and the PSMatching macro presented at SAS Global Forum 2007 by Marcelo Coca Perraillon. http://www2.sas.com/proceedings/forum2007/185-2007.pdf It assumes that you have already calculated a single numeric score variable that will be used to match cases to controls. Here is the program header that describes what it can do. %macro psmatch /*---------------------------------------------------------------------- Propensity Score Matching - match treated cases to controls ----------------------------------------------------------------------*/ (treated=       /* Input dataset of treated cases (REQ) */ ,control=       /* Input dataset of controls (REQ) */ ,out=           /* Output dataset of matched id numbers (DEF=MATCH) */ ,method=        /* Matching method (NN RADIUS CALIPER) (DEF=NN) */ ,caliper=       /* Pscore delta (used by CALIPER and RADIUS method)*/ ,id=            /* Name of ID variable (DEF=ID) */ ,pscore=        /* Name of Propensity Score variable (DEF=PSCORE) */ ,nmatch=        /* Number of controls to match per subject (DEF=1) */ ,replacement=   /* Select controls with replacement? (def=Y) */ ,seed1=         /* Seed for randomizing treated case order */ ,seed2=         /* Seed for randomizing control order */ ); /*---------------------------------------------------------------------- Propensity Score Matching Match treated cases to one or more controls. This does simple distance based on absolute difference between the Propensity Score of the case and the potential matched control. The propensity score variable must already be in the input datasets. Three methods of selecting the matches are available. Methods   NN      - Nearest neighbor. Find control with closest propensity score.   CALIPER - Same as NN but only consider scores within +/- &CALIPER   RADIUS  - Match ALL controls within +/- &CALIPER Based on ideas in paper by Craig Roberts and John R. Gerlach and the PSMatching macro presented at SAS Global Forum 2007 by Marcelo Coca Perraillon. http://www2.sas.com/proceedings/forum2007/185-2007.pdf Assumptions: - TREATED and CONTROL input datasets use same variable names for ID    and  propensity score variables. - The propensity score variable must be numeric. - The ID values must be unique keys for their datasets. Calls to: parmv.sas varexist.sas Outputs:   Dataset with list of matched cases and controls. The output dataset is not sorted. Unmatched treated cases are not included in the output.   VAR      Description   -------- -------------------------------------------------------   IDT      Treated case ID   IDC      Control case ID   DISTANCE Distance between the propensity scores (absolute value) Usage Notes: - When using SAS Version 9.1 or higher will use a hash object. - Creates work datasets _TREATED and _CONTROL. - When NMATCH > 1 then also creates and deletes view _TREATV_. - When METHOD=RADIUS replacement parameter does not apply. Controls    are selected within a treated case without replacement but are    available for matching to other treated cases (with replacement). - Treated and Control datasets are randomly ordered so that when there    are multiple controls of equal distance a random one will be    selected. ----------------------------------------------------------------------- History: 02AUG2009  abernt Creation ----------------------------------------------------------------------*/ ... View more

Code a new variable and sum that. data want ;   set have;   if involvement='event attendence' then wt=.5; else if involvement='leadership' then wt=2; run; ... View more

The SUM function will aggregate over the observations for given set of values of the grouping variables.  Normally you would want that OUTSIDE the case statement.  In this example the case statement will assign a number to each of the observations and then the SUM function will add them up. There might be situations where you would put the sum function inside the case statement, but this example does not look like one. ... View more