There are many syntax errors in your sql code. Compare it to next code: PROC SQL; Create table innerjoin_ as Select x.*, y.del_exc_rat, y.lagcogs, y.lagsel_gen_adm, y.lagope_lea_exp, y.lagr_and_d, y.laggoodwill from winsed.merge_treat_con as x inner join work._9blag as y on x.year=y.year and x.Type = y.Type; Quit; ... View more

1)  MID is not a sas function (origin probably from execl). Replace it with SUBSTR. The syntax is the same   substr(<string>,<from position>,<length>); If length is omitted it will extract substring up to the end of the string.   2) assuming ISSUE_DATE is a sas numeric date you need to use a specific format of date literaL:  20180930 should be '30sep2018'd. The literal of a time stamp (date and time) is '30sep2018:00:00:00'dt; A sas date is counting days since 01jun1960 which holds the value 0. (excel holds number of days since 01jan1900). ... View more

I adapted my code to your last explanation and answer to @Reeza  to get the wanted results: data want(drop=ID treat date rename=(svid=ID svtrt=treat)); set have end=eof; by ID notsorted treat; retain svid svtrt start; format start stop date9.; if _N_ = 1 or first.ID then do; svid = ID; svtrt = treat; start = date; end; else do; if treat=2 then do; if svtrt=1 then do; stop = date-1; output; end; svtrt=2; end; else svtrt=1; if treat=2 and last.treat then start=date+1; if first.treat then svtreat=treat; if last.ID then do; stop=date; if svtrt = 1 then output; end; end; run; proc sort data=want; by ID treat; run; ... View more

Assuming VISIT is a numeric sas date variable then do: proc sort data=jave; by ID visit; run; data want; set have; by ID visit; if last.ID; run; ... View more

If I understand you correctly you want the horizontal sum of subject per row in a report then check next codes: (1) summing all SUBs in a row: data students; infile datalines; input id sub1 sub2 sub3 sub4 sub5; total = sum(of sub:); datalines; 1 35 55 66 41 53 2 22 44 35 46 55 3 35 88 95 55 25 4 39 78 96 44 74 5 45 71 60 90 35 ; run; 2) Adding the total column in proc tabulate: proc tabulate data=students f=6. out=work.tabout; class id; var sub: total; tables id, ((sub:) total)*(N sum); run; 3) Adding the total column in proc means: proc means data=students sum maxdec=0; class id; var sub: total; run; 4) Adding the total column in proc report: proc report data=students nowd out=work.repout; column id sub1 sub2 sub2 sub3 sub4 sub5 total; define id / group style(column)=Header; define sub: / sum "sub Sum"; define total/ sum 'Total Sum'; run; I don't know if it is possible to show two different statistics of same variable(s).   If the out of above code is not what you want then post the format of the wanted report ... View more