I have a data set with a survey_identifier (school/class) which looks like 1_1_ for school 1, class 1 in which there will be many students. And I will have numerous schools with many class numbers. So for instance I will have 25 rows with 1_1_ because 25 students will be in that class. I am trying to use a Hash procedure to create separate data sets for each school but am getting an error "Uninitialized object at line 1030 column 16." Can someone help me figure out the issue. I have    DATA _NULL_;   declare hash eachschool (dataset: 'redcap1', ordered: 'Y');       eachschool.definekey ('redcap_survey_identifier', 'cn');       eachschool.definedata (all: 'YES');       eachschool.definedone (); declare hiter heach ('eachschool');   declare hash hashnum;   do until (done);      set redcap1 end = done;         if eachschool.check () then do;             hashnum = _new_ hash (ordered: "Y");             hashnum.definekey ('redcap_survey_identifier', 'cn');             hashnum.definedata (all: "YES");             hashnum.definedone ();      rc = eachschool.replace(); end;     rc=eachschool.find();     rc=hashnum.replace(); /*THIS IS THE LINE GIVING ME THE ERROR*/ end; do while (heach.next() = 0);     rc= hashnum.output (dataset: "school"||redcap_survey_identifier); end; stop; run; ... View more

ODS RTF FILE = 'S:\blahblah\frequencies.RTF' STYLE = BARRETTSBLUE STARTPAGE = NO; ODS NOPROCTITLE; %MACRO PRINTFREQ (VAR = ); PROC FREQ DATA =dat._3_subset; TABLES &VAR.; WHERE &VAR. NOT IN ('7', '8', '9', '77', '88', '99'); RUN; %MEND PRINTFREQ; %PRINTFREQ (VAR=Q1) %PRINTFREQ (VAR=Q2) %PRINTFREQ (VAR=Q3) I am using 9.4   X64_7PRO platform X64_7PRO WIN 6.1.7601 Service Pack 1 Workstation   As you can see in the attachment my output has two percent columns. Can someone tell me why I am getting two percent columns in the output and also why I am not getting a decimal in the percent column?  Is it a settings issue?  Can you assist me in resetting if needed?  Thanks. L ... View more