Thanks Frank;

Your Solution works perfectly!

And you are right it is a case where the "..the customer is always the customer."

( I offered every other format possible but they want excel). More work for them especially when the dates overlap files... but I warned them).

Thanks again! James

frank_diiorio@yahoo.com wrote: > The MIN and MAX values in the initial program reflect the date extremes > from the entire input dataset. If you want each "bucket" of the Excel > files to reflect the dates more accurately, you could do something like > the following - it takes the min and max dates for each group of > observations. Using the SUFFIX macro variable in the file name > prevents overwriting if more than one group of observations has similar > min and max values. > Why anyone would want to handle this amount of data in Excel is beyond > me, but as we all know from our days as retail trainees at Sears: "the > customer may not always be right, but the customer is always the > customer." > > data test; > do i = 1 to 11000; > eventDay = '01jan1950'd + i; > output; > end; > run; > > %macro cre8xls(data=, file=, size=5000); > > proc sql noprint; /* assume it's a WORK dataset. > If not, tweak the WHERE clause */ > select nobs > into :n > from dictionary.tables > where memname="%upcase(&data.)" & libname="WORK" > ; > quit; > > %let suffix = 0; > %do i = 1 %to &n. %by &size.; > %let suffix = %eval(&suffix. + 1); > %let endObs = %eval(&i.+ &size.-1); > > proc sql noprint; > select min(EventDay) format=date9., max(EventDay) > format=date9. > into :min, :max > from &data.(firstobs=&i. obs=&endobs.); > quit; > > %put Iteration &suffix. - obs &i. - &endObs. Min: &min. > Max: &max; > > proc export data=&data.(firstobs=&i. obs=&endObs.) > OUTFILE="c:\temp\&file. &suffix. from &min. to > &max..xls" > DBMS=EXCEL2000 REPLACE; > run; > %end; > %mend; > > options mprint; > %cre8xls(data=test, file=testxls, size=5000);