Sorry, I'm attempting this inside a data step, but I'm open to the most logical solution if it's not in the data step. Thanks. ... View more

Weird, my output changed from a number to this:   RULE_ORDER SCORE_IMPACTING RULE_NM TOTAL 999.3 Y TOTAL BIDS VOLUME SAMPLED co03a   Do I have the correct syntax for the Total variable?   226 ods tagsets.excelxp file="c:\users\ssbuechl\FinalData.xml" style=statistical options(embedded_titles='yes' 226! orientation='landscape' gridlines='yes' THOUSANDS_SEPARATOR=','); NOTE: Writing TAGSETS.EXCELXP Body file: c:\users\ssbuechl\FinalData.xml NOTE: This is the Excel XP tagset (Compatible with SAS 9.1.3 and above, v1.122, 01/04/2011). Add options(doc='help') to the ods statement for more information. 227 proc print noobs data=FinalData; 228 * format Rule_Order best6. total comma10. _: comma10. null comma10.; 229 var rule_order score_impacting rule_nm; 230 var total / style={tagattr='format:comma'}; 231 var _: null; 232 run; ... View more

Forum genius Tom helped me with this yesterday...   One method is to create a dummy set of records that has all of the dates in the order that you want so that PROC TRANSPOSE will see those dates first. Right now it is seeing them in the order that they appear after the data has first been sorted.  You can then drop that resulting row from the output dataset.   proc sort data=FinalData (keep=ad_dt) out=dates nodupkey; by ad_dt; run; data both ; set dates finalData ; run; proc transpose data=both out=Want(drop=_name_ where=(rule_nm ne ' ')); by rule_nm rule_order; id ad_dt_char; var n; run;   Proc transpose creates the columns as it sees new values of the ID variable. So by putting at least one example of every possible date in the first BY group (and in the right order) then it creates the columns in the right order.  Then the WHERE= dataset option deletes the extra group from the output.  Consider a  simple example.   data have ; input gender $ year count @@; cards;F 2003 10 F 2005 11 M 2002 12 M 2003 13;;;; proc transpose data=have out=want ; by gender ; id year ; var count; run; This will result in columns GENDER _2003 _2005 _2002. But if I add an extra set of records that before the F group that has YEAR in the order 2002,2003,2005 then the columns in the output dataset will be created in that order. ... View more

This is a great tip. Thanks so much for the help. ... View more

Thanks so much for the help. I appreciate your time. ... View more

How does setting Dates and FinalData together work exactly? Aren't the two datasets just appended to each other? I guess I'm not seeing how that works, but it appears to have the proper affect. ... View more

So is this physically rebuilding the FinalData dataset by creating the variables in the order of the 4 Set statements? ... View more

Can you explain the 'if 0' part of the code? I'm not sure what is 0? ... View more

They are a char type because I needed to have columns named 'Total' and 'Null' coming out of the Proc Transpose.  I also have the original ad_dt date variables available in the dataset, but the Total and Null columns are missing values.  This is why I coverted them to ad_dt_char and recoded them to 'Total' and 'Null'.   The ad_date_char values values are coverted in a previous datastep:   ad_dt_char=PUT(datepart(ad_dt), mmddyy10.); * Convert to a char value so we can recode missing values to text values of Null and Total ; ... View more

Now working outside the Rsubmit. Thanks to all for the help! ... View more

Does it matter that my program is running on the server-side within an Rsubmit block? Do I need to somehow bring the dataset FinalData to my local Windows machine? ... View more