Try:   Data A; Do i=1 To 18; Output; End; Run; Proc SQL NoPrint; * If a dataset is required use; Create Table xx As Select Nobs From Dictionary.Tables Where UpCase(Libname)='WORK' AND UpCase(Memname)='A' ; * If you need a macro-variable; Select Nobs Into :M_Nobs Trimmed From Dictionary.Tables Where UpCase(Libname)='WORK' AND UpCase(Memname)='A' ; Quit; %Put ***&M_Nobs.*; ... View more

It's not a question of the version, but of the module (if you type: proc setinit; run; somewhere in the log "SAS/OR" should appear).   I am afraid this is tricky. The only alternative I can think of is a fuzzy merge, but the results will be somewhat arbitrary - not an optimization. However, you could try this out with your actual data and take a look "how bad" the results actually are:   data cases; input ID $ wage; datalines; 1 730 2 1000 3 450 4 970 5 330 6 690 7 1750 9 1800 ; run; data candidates; input ID $ wage; datalines; A 700 B 690 C 720 D 730 E 1400 F 430 G 230 H 480 I 1390 ; run; Data _NULL_; cases=Open('cases'); Call SymputX('N_Max',Attrn(cases,'Nobs')); Run; %Put &N_Max.; %Let Const=1e9; %Macro Loop; %Do i=1 %To &N_Max.; Data Want&i. (Keep=ID_: best_:); If _N_ eq 1 Then Do; Declare Hash H (Dataset:'candidates',Ordered:'y'); H.Definekey('wage'); H.Definedata('wage','ID'); H.Definedone(); Declare Hiter HI ('H'); If 0 Then Set candidates; End; Set cases (Rename=(ID=ID_cases wage=ID_wage) Firstobs=&i. Obs=&i.); rc=HI.First(); Diff_Min=&Const.; Do While (not rc); Diff=Abs(Sum(ID_wage,-wage)); If Diff lt Diff_Min Then Do; Diff_Min=Diff; best_ID=ID; best_wage=wage; Call SymputX('Del_wage',wage); End; Else Do; Diff_Min=&Const.; Leave; End; rc=HI.Next(); End; Run; %Put &Del_wage.; %If &i.=1 %Then %Do; Data Want; Set Want&i.; Run; %End; %Else %Do; Proc Append Base=Want Data=Want&i.; %End; Data candidates; Set candidates (Where=(wage ne &Del_wage.)); Run; %End; %Mend; %Loop ... View more

The problem are the ties, aren't they. You might actually have to optimize ..   data cases; input ID wage; datalines; 1 800 2 1000 ; run; data candidates; input ID wage; datalines; 5 700 6 600 8 2000 ; run; Data Multi (Drop=rc); If _N_ eq 1 Then Do; Declare Hash H (Dataset:'candidates',Ordered:'y'); H.Definekey('wage'); H.Definedata('wage','ID'); H.Definedone(); Declare Hiter HI ('H'); If 0 Then Set candidates; End; Set cases (Rename=(ID=ID_cases wage=ID_wage)); rc=HI.First(); Do While (not rc); Diff=Abs(Sum(ID_wage,-wage)); Output; rc=HI.Next(); End; Run; Data Links; Set Multi (Keep=ID_Cases ID Diff); Rename ID_Cases=From; Rename ID=To; Rename Diff=Upper; Weight=1; Run; Data Nodes; Set cases (Keep=ID Rename=(ID=Node)); Weight=1; Output; Set candidates (Keep=ID Rename=(ID=Node)); Weight=-1; Output; Run; Proc OptNet LogLevel=moderate Graph_Direction=directed Data_Links=Links Data_Nodes=Nodes Out_Links=Matches (Keep=From To mcf_flow Where=(mcf_flow eq 1) Rename=(To=candidates From=cases)); MinCostFlow LogFreq=1; Run; Proc Print Data=Matches (Keep=c:); Run; ... View more

P.S. The Smiley isn't really a Smiley, but "Drop= :_ )". ... View more

I am not sure if I have fully understood what you are after, but I think:   Data A (Drop=i);   Format Date Date9.;   Do i=1 To 200;     Date=IntNX('day','01JAN2015'd,i-1,'same');  Value=Round(Rannor(1)*5+50,1);  Output;   End; Run; Proc TimeData Data=A Out=_NULL_ OutArray=A (Where=(not Missing(Value2)) Drop=_:);   ID Date Interval=Day;   Var Value;   OutArray Date2 Value2; * I don't think you can insert a date format for Date2 here .. ;   Do t=1 To _LENGTH_;     Value2[t]=Value[t-100];  Date2[t]=Date[t-100];   End; Run; Data A;   Set A;   Format Date2 Date9.; Run; ... View more