Thank you, Astounding! Your explanation is very helpful. Thank you very much on this one! ... View more

Thanks Peter.C! What I cannot understand is:  why the returned code of a macro, when this macro was invoked inside another macro, was masked (the returned semicolon was not treated as the end of a statement). If we add another %PUT statement in the macro A: Program %macro a;     part1; part2     %put This is when macro A execue.; %mend; %macro b;    %put *%a*; %mend; *** Try 1; %b *** Try 2; %put *%a*; Log 19   %b This is when macro A execue *part1; part2* 20   %put *%a*; *part1 This is when macro A execue NOTE: Line generated by the invoked macro "A". 1    part1; part2             -----             180 ERROR 180-322: Statement is not valid or it is used out of proper order. We can see from the log that when macro B executes, the execution of macro A was before B's %PUT statement. So I am guessing maybe there's an automatic quoting when a macro was invoked inside another macro definition. But I am not sure whether this guess is right or not. Thanks again! ... View more

Thanks Amir! I have the same thought too, but didn't find any proof in the Help. ... View more

Hi Urvish, You are right, most of the times GROUPFORMAT would be used with SET statement.  Actually I used GROUPFORMAT for only a few times. I ran into this issue when trying to answer someone else's question about merging data sets. While merging, variables from data sets would not be set to missing at the beginning of each iteration of data step. Of course, at the beginning of the first iteration, all variables would be initialized as missing when no data has been read from data sets. Thank you very much for spending time on this! ... View more

Hi Urvish, I am using SAS 9.1.3 and I didn't get any ERROR message on this example. Thanks! ... View more

Hi Urvish: Good morning and thanks for your reply. We know that in sashelp.class the value is "Jeffrey" for NAME. The purpose of using variable NAME1 is to store the unformatted value of name, and show in the log which value ('jEFFREY') can be successfully merged to 'Jeffrey' and which ('JEFFREY') cannot. name=JEFFREY n=1 name1=JEFFREY Sex=  Age=. Height=. Weight=. FIRST.name=1 LAST.name=1 _ERROR_=0 _N_=6 name=JEFFREY n=2 name1=jEFFREY Sex=M Age=13 Height=62.5 Weight=84 FIRST.name=1 LAST.name=1 _ERROR_=0 _N_=10 I expected both values can be merged to "Jeffrey" when using GROUPFORMAT + FORMAT, but it turned out wrong.  So I am not looking for a way to merge datasets, I am looking for a reason why this method didn't work. Thanks! ... View more

Yeah, this is totally beyond my understanding. BTW, there will be no ERROR if you add a 'd' in your "data one;". data one ;   input id $ @@ ;   copy = id ;   format id $upcase. ;   put (_all_) (=); cards; a  B b c d D e run; Thanks again! ... View more

Thanks Tom！ But please try the following example. All observations can merge successfully and no complain about the alphabetic order. Thanks for your time! data test;      input name $8.;   n=_n_;   cards; henry HENRY Henry HENRY HenRY HENry jeffrey THOMAS ; data test1;      format name $upcase.;      merge test sashelp.class(in=okay);      by name groupformat;      if not okay then put "Observation " n "is not okay!" name= $8.; run; ... View more

Thank you Urvish! And I know we can use PROC SQL and some other ways to do the merge. My real purpose is not to merge two data sets to get a certain result. My purpose is to get to know why the GROUPFORMAT + FORMAT failed to generate the expected result. There should be something that I didn't correctly understand. $UPCASE is an internal format in SAS. Thanks again! ... View more

Thanks for your response, but it's not because of the FORMAT statement. Log: 4632  data test1; 4633     format name $upcase.; 4634     merge test sashelp.class; 4635     by name groupformat; 4636     if upcase(name)="JEFFREY" then put _all_; 4637  run; name=JEFFREY n=1 name1=JEFFREY Sex=  Age=. Height=. Weight=. FIRST.name=1 LAST.name=1 _ERROR_=0 _N_=6 name=JEFFREY n=2 name1=jEFFREY Sex=M Age=13 Height=62.5 Weight=84 FIRST.name=1 LAST.name=1 _ERROR_=0 _N_=10 NOTE: There were 2 observations read from the data set WORK.TEST. NOTE: There were 19 observations read from the data set SASHELP.CLASS. NOTE: The data set WORK.TEST1 has 20 observations and 7 variables. NOTE: DATA statement used (Total process time):       real time           0.08 seconds ... View more