Can you explain more what is the process that is generating this data?   Looks like the pattern is that when you have a missing value of PREVSUBJ you want to find the largest value of SUBJ in that block.  Might be easiest to first generate the group numbers. data have; input subj prevsubj; cards; 1001 . 1002 1001 1003 1002 1011 . 1012 1011 2010 . 2011 2010 2012 2011 2013 2012 2014 2013 ; data step1; set have ; group + missing(prevsubj); run; proc sql; create table want as select group, subj, prevsubj, max(subj) as newsubj from step1 group by group order by group, subj ; quit; Result Obs group subj prevsubj newsubj 1 1 1001 . 1003 2 1 1002 1001 1003 3 1 1003 1002 1003 4 2 1011 . 1012 5 2 1012 1011 1012 6 3 2010 . 2014 7 3 2011 2010 2014 8 3 2012 2011 2014 9 3 2013 2012 2014 10 3 2014 2013 2014   ... View more

Thank You for the help Cynthia!   I opened a track with Tech Support Team. ... View more

Hi @SR2019    Linux SAS cannot handle native Excel file formats, only xlsx-files, and there is no OLE support. But if your spreadsheet is xlsx, you can read xlsx files with proc import or the xlsx libname engine and write them proc export, the xlsx libname engine or ODS the same way as in windows. There is no difference.   Usually, the difficult part is to access the files, because spreadsheets normally resides on windows machines in places where the linux system can't see them, so it is necessary to FTP them to linux or move them to a windows folder, that is mounted on the linux system.    If it is not meant for production there is also SAS Enterprise Guide, and for production work there is a SAS PC files server that might make things a little easier, but I have no experience with it and don't miss it, because both FTP and mount works well for us, at least as well as the match between SAS and Excel usually works. These two are not meant for each other!     ... View more