For you 12:19 post, could you perhaps post an example of the output you want.  Will take a look tomorrow. As for the other post, doing the same code multiple times can be done via macro or call execute.  E.g. ods tagsets.excelxp file="...afile.xlsx"; data _null_;      array ds{2} $200. ("SASHELP.CARS","SASHELP.CLASS");      do i=1 to 2;           call execute(' ods tagsets.excelxp options=(sheet_name="'||strip(ds{i})||'");                               proc report data='||strip(ds{i})||'; run;');      end; run; The above code will produce some code to name an output sheet, then report the data from the dataset out.  So you comments could be an empty dataset that you output as a new sheet... ... View more

Ok, its even easier then: proc sql;      create table RESULT     /* Create an empty results table which we will populate by the statements further below */      (           REC_SUMMARY char(200),           FREQ char(200),           PCENT char(200)      );      insert into RESULT      set     REC_SUMMARY="Match",          /* Here we start inserting the first row of the results which is those that match */                FREQ=(select put(count(*),best.) from HAVE where COMMENT="MATCH"),   /* Count all records with MATCH */                PCENT=put( (select count(*) from HAVE where COMMENT="MATCH") / (select count(*) from HAVE) *100,5.2)   /* Num rows with Match / all rows * 100 = percent */      set     REC_SUMMARY="Mismatch reason 1",               /* Do same as above but using condition doe reason 1 - which I can't identify from what you have given */                FREQ=(select put(count(*),best.) from OUT_A)...      set     REC_SUMMARY="Start Date",     /* Insert the min start date */                FREQ=(select min(POLDTR) from OUT_BOTH)      set     REC_SUMMARY="End Date",     /* Insert the max start date */                FREQ=(select max(POLDTR) from OUT_BOTH); quit; ... View more

Well, I have just finished for the day, but thought I would throw out a couple of ideas (note this is is in no way complete, just gives a flavor.  Would need to see some test data and what reason 1/2 is and what constitutes a match): data out_both out_a out_b;      merge data1 (in=a) data2 (in=b);      by by_vars;      if a and b then out_both;      else if a then out_a;      else out_b; run; proc sql;      create table RESULT      (           REC_SUMMARY char(200),           FREQ char(200),           PCENT char(200)      );      insert into RESULT      set     REC_SUMMARY="Match",                FREQ=(select put(count(*),best.) from OUT_BOTH),                PCENT=(select put( (count(*) / &TOTAL.) *100,5.2)                               from OUT_BOTH)      set     REC_SUMMARY="Mismatch reason 1",                FREQ=(select put(count(*),best.) from OUT_A)...      set     REC_SUMMARY="Start Date",                FREQ=(select min(DATE) from OUT_BOTH group by ..)... quit; ... View more

Hi, that seems to be an example from the help page.  Anyways, use the of colx-coly.  Note it will trim your data and hence lose the spaces. data _null_;    col1='The Olympic  ';    col2='   Arts Festival ';    col3='   includes works by ';    col4='Dale Chihuly.';    result=cat(of col1-col4);    put result $char.; run; If you need the $ then put it into your strings. ... View more

If you do it that way, you are very quickly going to run out of column width.  Remember there are page width maximums, and if you are going to Excel then column maximums.  You could do it in a datastep call execute: data _null_;      call execute('proc report.....;');      do I='01JAN2014'd to today();           call execute(' define '||put(I,date9.)||' / "'||put(I,date9.)||'";');      end; run; That will create a define line for each date between 01Jan and now, so something along those lines.  However, again, I would consider re-organising your data/output in a different way. As for how to create a title page, well that is a whole other story.  There are several topics on this.  Its dependant on your output type, what you want in a contents etc.  Its definitely not a short topic, so please search these forums, also a paper: http://www.nesug.org/Proceedings/nesug06/io/io13.pdf ... View more

Not sure what the column vector array means, but with normal arrays you could use if statements, select/when statements, in fact any conditional logic in a datastep or sql to change the information e.g.: if the_array{iterator}=0 then 0; if the_array{iterator}=1 then 1; ... if the_array{iterator} in (14,15,16,17,18,19,20) then 7; ... View more

Not sure I am following what the problem is.  You could change the code slightly to: proc sql;      create table LOOP as      select distinct COUNTRY from HAVE; quit; ods tagsets.excelxp file="...\file.xlsx"; /* This loops over each country and prints the dataset with the corresponding country name to the tagset */ data _null_;      set LOOP;      call execute('ods tagsets.excelxp options=(sheet_name="'||strip(country)||'");                           proc print data='||strip(country)||';                            run;'); run; ods tagsets.excelxp options=(sheet_name="Help"); ... ods tagsets.excelxp close; ... View more

Yep, MS have adopted that format.  Its ok, just zipped XML.  You are correct though, looking back over the ExcelXP tagset definition there is something called flyover which can do comments.  Have a look at the documentation: Base SAS: Demo: ExcelXP Tagset and Microsoft Excel Not sure I would want to go that route myself, however it does look possible. ... View more