I suck at code but if I interpret your question correctly you can sort your data and create a new column using a 'first' function and count the rows within a specific group... sort your data descending (last color on top) count per group with a rank number +1 for every new instance until you encounter the next group (The K-11,K-12 in your data set I think?) And then only filter the rows labelled 1 till 8 and reverse them again if needed. data "have";      SET "output tabel"; by "K-11 colunm name"; RANK +1;      if first."K-11 colum name" then RANK = 1; RUN; first sort it the right way, and maybe add other column to your BY group when needed. ... View more

Yeah this is a strange one.   First if this is your source table and you are querying it multiple times and there are duplicate filters in those query builders you called 'filter' then you might want to add a step next time. I make sure all my source tables have at least 1 query builder (sometimes even just a copy, no filters) before using it multiple times. The benefits are. 1. When something like this happens you can just query your new source table and rename the output table to the excisting one. For instance. name the output name from query builder B to the output name of query builder A. And then delete query builder A and the old source. 2. When you want to change the time period you only have to do it in query builder A. Also any new filters for both QB 1 and 2 can be done in one place. 3. Your source table is only approached once. (Benefits depend on the way your source tables are structured of course.) 4. When a new input table has different column names and you don't want to change them in multiple query builders or throughout every step in your query, just change them to the old ones in the first step.   Last but not least. There is also a fix for you problem I think. But it's in a weird place and you have to do it separate in all of your 'filter' query builders now. 1. open the join screen in your query builder (doesn't matter if your have an actual join or not). 2. Right click the name (or outlines) of your source table and choose properties 3. Source: --> Change en choose your new source table. This works most of the times but not always, I don't know why. At least make sure the data is available, and make a new query builder from your new source and connect to that one instead of to your source tabel directly! That way you can also make sure the column names are the same etc. Good luck! ... View more

140 views but no reply's. I understand that no one has the solution, but I am guessing no one has a problem with this behavior as well?. Or atleast does not share my frustration. Maybe I'll contact SAS directly and suggest it as a bug. ... View more

I don't know if this has anything to do with it but seeing there are no reactions, an I recently had some trouble with imported files in EG 8.3, why not give it a try.   In the properties of the imported file, what SAS server and Library do you use? I had the problem that my SAS server went tot 'local' automatically for imported files. And that gave some trouble because we all work in the WORK lib on the SASApp server. Maybe switch something up with the file storage location and see if the problem still exists. (I ended up hiding the whole 'local server')   If I didn't understand correctly and you are using files on the server itself, try 'opening' (unfold?) the directory untill you see the table in the server pane. Then run.   That's all I can suggest, good luck! ... View more