Bruce, thank you for posting!   All the responders, Thank you for educating me! %macro test(n); %if &n=QTR %then MSRP; %if &n=MONTH %then INVOICE; %mend test; %LET version = QTR;  *%LET version=MONTH; proc sql;   create table cars_total as   select distinct sum(%unquote(%test(&VERSION.))) as &version._%unquote(%test(&VERSION.))_total, Make    from sashelp.cars    group by make    ;   quit;   proc print data=cars_total(obs=9);run; ... View more

What output values are you expecting to see in variable V1?  Does V1 need to contain a character value representing the name of a variable or does V1 need to contain the value of a variable name? Kurts code: %macro test(n); %global varname; %if &n=1 %then %let varname=a; %if &n=2 %then %let varname=b; %mend test; data test; azz = 0; bzz = 1; %test(1); v1 = &varname.zz; %test(2); v2 = &varname.zz; run; proc print data=test(obs=7); run;                                         Obs    azz    bzz    v1    v2                                          1      0      1      0     1 XIA's code: %macro test(n); %if &n=1 %then  a ; %if &n=2 %then b ; %mend test; %put %test(1)zz; /* produces azz */ data x; azz=0; v1 ="%test(1)zz"; run; proc print data=x(obs=7); run;                                               Obs    azz    v1                                                1      0     azz %macro test(n); %if &n=1 %then a; %if &n=2 %then b; %mend test; %put %test(1)zz; /* produces azz */ data data4; azz=0; v1 = %unquote(%test(1)zz);  /*does not produce an error. V1 contains the value of variable AZZ */ run; proc print data=data4;run;                                                Obs    azz    v1                                                 1      0      0 ... View more

I created dataa and datab datasets changing 31JUN2014 to 30JUN2014.  The only problem I had was with ID 235.  Your 'want' says 235 should be gold.  My results say it should be SILVER. I arbitrarily gave days the value of 1000000 when it would calculate as missing. proc sql; drop table testit; create table testit as select distinct  a. ID, a.date format=date9., case when b.id  = ' ' then . else b.date2 end as date2 format=date9., case when b.id = ' ' then ' ' else b.membership end as MEMBERSHIP, case when b.id = ' ' then 1000000 else a.DATE-b.DATE2 end as days , abs(calculated days) as absdays , min(calculated absdays)as close from DataA as a full join DataB as b on  a.ID=b.ID  group by a.ID, a.date order by id, close, date, absdays, date2 ; quit; proc sort data=testit; by id date close absdays; run; data testit3; set testit; by id date close absdays; if first.date then output; else delete; keep id date date2 membership; run; proc print data=testit3;run;   ID      DATE              date2    MEMBERSHIP    days    absdays    close    235    31MAR2014    28FEB2014      SILVER        31       31        31    235    31MAR2014    31MAY2014      GOLD         -61       61        31                               Obs    ID          DATE        date2    MEMBERSHIP                                 1    123    31JAN2014    28FEB2014      PURPLE                                 2    123    31JUL2014    31JUL2014      SILVER                                 3    123    31AUG2014    31JUL2014      SILVER                                 4    222    31MAR2014    31MAR2014      BRONZE                                 5    222    30APR2014    31MAR2014      BRONZE                                 6    235    31MAR2014    28FEB2014      SILVER                                 7    333    30JUN2014            .                                 8    421    28FEB2014            .                                 9    991    31MAY2014            .                                10    991    31OCT2014            .                                11    991    31DEC2014            . ... View more

If this helps, SAS/ETS(R) 9.2 User's Guide   . "Note that the %DFTEST macro has been superseded by the PROC ARIMA stationarity tests. See Chapter 7, The ARIMA Procedure, for details." ... View more

This might help you.    I don't have the answer to your question.  I'm monitoring your question so that I can learn from your question. ... View more

I am using 9.4 64 bit and I still receive a '1' for case.  Can you give a stronger hint? proc sql; drop table test; create table test  as select d.region, case         when d.region in (select d.region from sashelp.bmt ) then  '1'           else                                                                         '0'         end  as case from sashelp.demographics as d ; quit;   Obs    region    case                                               1     AMR       1                                               2     AMR       1                                               3     AMR       1                                               4     AMR       1                                               5     AMR       1                                               6     AMR       1                                               7     AMR       1 ... View more