This doesn't work for a large dataset, you receive an out of memory error. ... View more

It is not a good idea to put data into metadata.  Any way you can have the Excel file transposed BEFORE you import it so the dates are in a column instead of being used as column headers?   There are ways to deal with trying to read the dates as data using PROC IMPORT by adding the GETNAMES=NO optional statement.  Read the dates line as one dataset and the body of the data as another.   But your real question is how to use the date value you do have as a variable so that you can find the values that correspond to the follow-up months.   So let's assume you have this data set. data have ; input Emp_id Joining_date _2016_04_01 _2016_05_01 _2016_06_01 ; informat joining_date yymmdd. ; format joining_date yymmdd10. ; cards; 101 2016-04-05 0.15 0.40 0.28 102 2016-04-21 0.56 0.34 0.67 103 2016-05-13 0.39 0.61 0.84 ;;;; Note: Don't use DMY (or MDY) order to display dates. Either way you will confuse half of your audience.   So now to make new variables that have the values relative to Joining_date you just need calculate the indexes into the two arrays. data want ; set have ; array absolute _201: ; array relative (3) ; do from=intck('month','01APR2016'd,joining_date)+1 to dim(absolute); to=sum(to,1); relative(to) = absolute(from); end; drop from to; run; Joining_ _2016_ _2016_ _2016_ Obs Emp_id date 04_01 05_01 06_01 relative1 relative2 relative3 1 101 2016-04-05 0.15 0.40 0.28 0.15 0.40 0.28 2 102 2016-04-21 0.56 0.34 0.67 0.56 0.34 0.67 3 103 2016-05-13 0.39 0.61 0.84 0.61 0.84 . ... View more

Firstly use PROC PRINTTO; to save LOG information and then use data step to count them all. proc printto log='c:\temp\saslog.log' newfile; run; ..........your sas code here............. proc printto;run; data _null_; infile 'c:\temp\saslog.log' end=last; input; if _infile_ =: 'ERROR' then count+1; if last then putlog 'The number of ERROR is ' count ; run; ... View more