On 5/1/08, Theorbo <reply@to-group.nfo> wrote: > Hi. I have a dataset like this: > ID, MONTH, DAY, MeasureA, MeasureB, MeasureC > 1,2,13,243,221,254 > 2,2,13,244,255,263 > 1,2,13,255,278,233 > 1,2,14,223,211,266 > 2,2,15,223,277,265 > > with multiple observations per ID, month and day. I have about 25 of these > measures - all numeric. > > I am trying to get to this output dataset (to dump into Excel - don't get me > started!): > > ID, Avg_A_Month1, Avg_B_Month1, Avg_C_Month1, Avg_A_Month2, Avg_B_Month2 > > where I have one observation per ID with averages for the measures > aggregated by month. > > I know it's not good to keep my "data" in the variable names - isn't that > right? > What's the best way for me to get from start to output? > I started by trying to do a proc means nway noprint, class month, out= ... > and then transposing the output dataset but then I'm left with one variable > that contains all the variable names. Do I need to transpose twice? > > Any ideas? Thanks for any tips or suggestions - no code necessary.

The MEANS procedure wrapped inside the HTML output of ODS, which can be directly opened in Excel and saved as a .xls file?