Thank you @vidyasagar1. You're welcome ... View more

Hi @IdlanHnf    As an alternative to PORC SQL, you can also use a simple PROC MEANS:   proc means data=myData nway noprint; class buyer; var price; weight unit; output out=buyer_totals (drop=_:) sum= / autoname; run; If you need a report -> remove the NOPRINT option If you need a table -> keep the OUTPUT OUT= statement. Best,   ... View more

Hi @Ronein    What if you add the following code to the TABLES statement? Table z1=""*z2='' all, /*added:all*/ z3='Nr customers'*z4*N all='income'*N='Grand total N' z3='Sum Income'*z4*y all='Sum Income'*y='Grand total income' z3*N /*added*/ z3*y /*added*/ Best, ... View more

Hi @lillianlui    I am not sure there is a simple answer to your question. ORDER=FREQ works well for a one-way frequency table but it seems tricky to get similar results with a two-way frequency table. Here is an approach to get the desired results: 1/ Output the results of PROC FREQ in a dataset (here it is work.np_codelookup) 2/ Sort the results according to the type and the count 3/ Create the report based on this output.   /*** Run a proc freq and output the results in a dataset: work.np_codelookup */ proc freq data=pg1.np_codelookup order=freq noprint; tables type*region / outpct out=work.np_codelookup (drop=percent pct_col); where type in ('National Historic Site', 'National Monument','National Park'); run; /*** Sort the results by type and descending counts */ proc sort data=work.np_codelookup; by type descending count; run; /*** Create the report */ /* NB: TIP: - the 'group' category in the define statement for type and region allows you to suppress repetitious printing of the values - the break after statement allows you to compute the "ALL" row for each type - creating the new 'region_dummy' variable allow you to add the 'ALL' in the region field */ proc report data=work.np_codelookup_weight; column type region region_dummy count pct_row; define type / group; define region / group noprint; define region_dummy / computed 'Region'; define count / analysis; define pct_row / analysis f=8.2; break after type / summarize suppress; compute region_dummy / char length=20; if _break_='Type' then region_dummy="Total"; else region_dummy = region; endcomp; run; Best,     ... View more

Hi @cmemtsa  You cannot update the T table by using an inner query that uses this table as it is already in use. I would suggest that you create another table: proc sql; create table T2 as select *, sum(M) as M1 from T group by ID, L; quit; Best, ... View more

Thank you so much @Rentum ! ... View more

You’re welcome 😊  ... View more

Hi @Caro17  Please try this, to begin with the character in first position diag = substr(diagnose,1,3); Best, ... View more