One way I think of is;   Use the proc means code and then do a    proc sql; create temp table as  select year, count(*) as tot from demo group by year;   data final; merge meanstable temp ; by year; run;   or drop columns in the earlier code  ... View more

I think OP wants overall count per year, but this code will count for all T1-T100, which will be basically same but will repeated 100 times like T1_N - T100_N, generally all he wants one column of count per year ... View more

your question is very broad and is very domain related question, I fear it will be very difficult to answer for anyone to answer. If you want to know more about EPOCH and SAS  then please use this link https://lexjansen.com/search/searchresults.php?q=epoch   If you want any specific answer for your data, please give a sample input data and output data, so that someone will able to help you. This way you will get your solution much quicker way   ... View more

one of the answers on SAS communities and other important links giving some info about this   https://communities.sas.com/t5/Administration-and-Deployment/CLI-error-connecting-to-SQL-Server/m-p/410188#M11046 http://support.sas.com/kb/19/707.html http://support.sas.com/kb/14/386.html ... View more

please  use explicit pass through to optimize query and example is shown below and also add fastexport = yes in your connect statement    proc sql; connect to teradata(server=server user=userid pwd=password fastexport =yes); create table work.job204 as select * from connection to teradata (select * from employees where jobcode=204 and name = 'smith'); quit; ... View more

i liked all the answers which are simpler and much cleaner, just one more way    data want; set have; IF _N_ =1 THEN PATTERN = PRXPARSE ("/((?<=ORIG:)\S+)/"); retain pattern; if prxmatch(pattern, orig) then call prxposn(pattern, 1, position, length); val = substr(orig, position,length); drop pattern position; run;     ... View more

long and not very efficient way. use this if you have few columns and few values data a; input value1 value2 value3; cards; 4 6 5 5 7 3 3 2 2 2 4 1 ; run; data b; input Parameter $ amount; cards; value1 .5 value2 .1 value3 1 ; run; data a1; set a; col =_n_; run; proc sql; select coalesce(tot1,0) +coalesce(tot2,0)+coalesce(tot3,0) as final_total from (select value1, col, case when value1 and parameter = 'value1' then value1*amount end as tot1 from a1 cross join b where calculated tot1 is not missing)x inner join (select value2,col, case when value2 and parameter = 'value2' then value2*amount end as tot2 from a1 cross join b where calculated tot2 is not missing)y on x.col =y.col inner join (select value3,col, case when value3 and parameter = 'value3' then value3*amount end as tot3 from a1 cross join b where calculated tot3 is not missing)z on x.col =z.col; ; ... View more

if you can give an example( 5 to 10 rows) input dataset and output dataset, it will be very easy for help.       ... View more

one way to do this data have; input @1 ID @5 Count @15 FY; datalines; 1 1 10 1 1 11 1 3 12 1 1 13 1 12 14 1 2 15 2 2 11 3 1 11 3 7 12 3 2 13 3 1 14 ; proc sql; create table want as select a.* from (select * from have)a inner join (select id, min(FY) as min_fy from have where count >= 2 group by id)b on a.id = b.id and a.fy >= b.min_fy;   ... View more