Curious if you attempted using PROC TRANSPOSE, after reading the DOC, and considering the examples that are available at the SAS support http://support.sas.com/ either using the SEARCH facility at the site or with an Internet search?
Also, you show 1 row for 20100120, then multiple rows for another date-related variable -- yet you attempt to describe the desired result which doesn't necessarily jive with the "input source".
Yes, That's correct. I have data for HRA and we get data by Completion date and by each of the question code followed by an answer in one record. I wanted to combine all of the unique questions related to that person into one record for all possible questions and I am not sure if PROC TRANSPOSE will do that.
As suggested, the SAS support http://support.sas.com/ website and its SEARCH facility can reveal much information for PROC TRANSPOSE - both SAS-hosted documentation and also supplemental conference/technical papers on the topic.
Also, the forum archives will provide useful reference information as well, when searched.
Message was edited by: sbb
Strongly suggest that you start with some instream DATALINES, or consider SASHELP.CLASS, so you can get familiar with the PROC TRANSPOSE processing.
I very much expect that you can emulate your output rqmts to-start with using SASHELP.CLASS, get the processing well-defined, then move forward (slowly) with introducing your specific data.
You need to learn the procedure/process first, then adapt the process(ing) to your specific input data conditions and output data requirements -- one step at a time.
And, remember, you know your data - we don't - so to reveal SAS code in your post/reply for which we have no clue how you mangled the input side is less-than-useful if you actually expect to get assistance from other subscribers (without making it a guessing contest).