Here is my try by using macro... %macro test;    proc sql noprint;       select distinct class into :list separated by "?"       from have;       select count(distinct class) into :tot       from have;    quit;    %do i = 1 %to &tot.;       data data&i.;         set have;         if class = %scan(&list.,&i.,"?");       run;    %end; %mend; %test -Urvish ... View more

Well, there are different ways to manipulate the data so i thought ARRAY will help to get the desired output data to be use for T-Test... Your code is also correct...Based on your very first post i was not that much clear about what you want to compare so code in such a way that will give split the data in two... Anyways you can also modified your above code so that user can pass any ANALYSIS VARIABLES... %macro t-test(analysis_var);    proc ttest data = have;       class drug1_drug2;       var &analysis_var.;    run; %mend; %ttest(discharge); -Urvish ... View more

Hi, If you want to make the infrence between Drug1 yes-Drug2 no. VS Drug1no-Drug2yes then PROC TTEST is suitable... Please have a look at following modified code... i have appended the two categories along with their averages and then calculated T-Test between them that Two Sample T-Test...Hope it meet the requirement data have;    input drug1 $ drug2 $ average;    cards4; yes no 9 no yes 13 . no 8 ;;;; /* Drug1(Yes) VS Drug2(No) */ data drug1_yes_drug2_no(drop = drug1 drug2);    retain drug average;    set have;    if drug1 = "yes" and drug2 = "no";    array temp(1) $ drug;    temp(1) = drug1;    output;    temp(1) = drug2;    output; run; /* Drug1(No) VS Drug2(Yes) */ data drug1_no_drug2_yes(drop = drug1 drug2);    retain drug average;    set have;    if drug1 = "no" and drug2 = "yes";    array temp(1) $ drug;    temp(1) = drug1;    output;    temp(1) = drug2;    output; run; data want;    set drug1_yes_drug2_no drug1_no_drug2_yes; run; proc ttest data = want;    class drug;    var average; run; -Urvish ... View more

INFORMAT 6.2 you can apply... ... View more

Hi, Here is my try...based on my understanding, it is the case of comparing two proportions of a categorical variable(CHI-SQUARE TEST)...so all you need to do is to manipulate the data before you use it to make infrences...The following one is the possible solution of your only 1 Question which you have posted...If it is meeting your requirement then we can try with second quetion... /* Drug1(Yes) VS Drug2(No) */ data drug1_yes_drug2_no(drop = drug1 drug2);    retain drug average;    set have;    if drug1 = "yes" and drug2 = "no";    array temp(1) $ drug;    temp(1) = drug1;    output;    temp(1) = drug2;    output; run; proc freq data = drug1_yes_drug2_no;    tables drug / chisq; run; /* Drug1(No) VS Drug2(Yes) */ data drug1_no_drug2_yes(drop = drug1 drug2);    retain drug average;    set have;    if drug1 = "no" and drug2 = "yes";    array temp(1) $ drug;    temp(1) = drug1;    output;    temp(1) = drug2;    output; run; proc freq data = drug1_no_drug2_yes;    tables drug / chisq; run; -Urvish ... View more

How about following one... proc sql;    create table want as    select tax_id,max(npi) as max,min(npi) as min,           case                   when calculated max EQ calculated min then "one_2_one"                     else "one_2_many"               end as relation    from have as a    group by tax_id; quit; proc sql;    create table one_2_one as    select a.*    from have as a, want as b    where a.tax_id = b.tax_id and          relation = "one_2_one";    create table one_2_many as    select a.*    from have as a, want as b    where a.tax_id = b.tax_id and          relation = "one_2_many"; quit; -Urvish ... View more

I know it has been marked as ANSWERED...but still here is my try if you like... data want;    set have;    i = 1;    str = scan(string,i);    do while(str NE " ");       j + 1;       _str_ = str;         i = i + 1;         str = scan(string,i);         output;    end; run; proc sort data = want;    by string indicator; run; proc freq data = want noprint;    tables _str_ / out = freq(drop = percent);    by string  indicator; run; proc transpose data = freq out = want(drop = _name_ _label_);    by string indicator;    var count;    id _str_; run; data want(drop = i);    set want;    array temp(*) _numeric_;    do i = 1 to dim(temp);      if temp(i) = . then temp(i) = 0;    end; run; -Urvish ... View more

Here is my try... proc sql;   create table want as   select id,year,          sum(value1) as value1,              sum(value2) as value2,              sum(value3) as value3   from have   group by 1,2; quit; -Urvish ... View more

Hi, How about the following one... proc transpose data = have                out  = have(rename = (_NAME_ = Month COL1 = VAL));       by id;       var month1-month12; run; data want;    set have;    by id;    _temp = dif(val);    if _temp = . then _temp = 0;    if first.id or _temp = 1 then con_mnth = 1;    else con_mnth + 1;    if _temp < 0 then con_mnth = .; run; proc sql;    create table want as    select id,max(con_mnth) as con_mnth    from want    where val = 1    group by id; quit; -Urvish ... View more

Hi, By Default, CALL SYMPUT will maintain any leading or trailling blanks...In below example program, i have created name which have 20 bytes length but originally there is only 3 bytes so name will have 17 extra trailling blanks which is going to retain while creating macro variable by using CALL SYMPUT during data step execution... But if you use CALL SYMPUTX insted of CALL SYMPUT then it will remove any leading or trailling blanks from the variable...Here, CALL SYMPUT "X" represents EXTRA BLANKS to be remove and in contrast, PROC SQL will retain any leading or trailing blanks while creating macro variable during proc sql execution... SEE THE BELOW EXAMPLE PROGRAME...HOPE IT HELPS...   data test;    length name $20.;   name = "XYZ";     call symputx("data_X_name",name);     call symput("data_name",name); run; proc sql noprint;     select name into :sql_name     from test; quit; %put CALL SYMPUT will resolve to: &data_name.*****; %put CALL SYMPUTX will resolve to: &data_x_name.*****; %put PROC SQL will resolve to: &sql_name.*****; -Urvish ... View more

Hi, Hope it works...In below macro no_of_vars parameter represents matching columns to be compare...like aa1,aa2,aa3, and bb1,bb2,bb3 and so on...so pass the value accordingly... Here you only have 1 common column like aa1 and bb1.... %macro obs_identify(no_of_vars =);     proc contents data = one noprint                     out  = one_list(keep = name where = (name ^in ("Cusip" "Year")));     run;     proc contents data = two noprint                     out  = two_list(keep = name where = (name ^in ("Cusip" "Year")));     run;       proc sql noprint;          select name into :one_list separated by ","          from both_list;          select two_name into :two_list separated by ","          from both_list;       quit;     proc sql;         create table both as         select coalesce(a.cusip,b.cusip) as cusip,                coalesce(a.year,b.year) as year,                &one_list.,&two_list.                case                    when a.cusip = b.cusip then "Match"                    when aa = . and bb = . then "Match"                            /* Loop for Series of Matching Columns */                            %do i = 1 %to &no_of_vars.;                        when bb&i. = . then "Two"                        when aa&i. = . then "One"                          %end;                    else "Match"                end as source         from one as a              full join              two as b         on a.cusip = b.cusip;     quit; %mend; %obs_identify(no_of_vars = 1); -Urvish ... View more