So with example data we can begin to get an idea what you are trying to do.  Looks like you have grouping variables (in this case 3 of them: Make Drivetrain Type) that uniquely identify the groups.   And you have a CLASS variable (in this case: Model) to subdivide the groups.  And an analysis variable (in this case MSRP) that you want to SUM.   One obvious savings is to use the output of the first summary (at the BY GROUP* CLASS VARIABLE level) to calculate the BY GROUP sums.   In your example that would be changing the step that makes CARS_MEANS to use CARS_FREQ as the input. proc summary data=cars_freq ; by make drivetrain type; var count ; output out=cars_means(drop=_:) sum=MSRP_sum_carstate; run; It just doesn't attach the DOLLAR format to the variable.   Here is how you can use one PROC SUMMARY (MEANS and SUMMARY are the same procedure, they just use different defaults for the PRINT/NOPRINT option) and one data step to calculate your COUNT, PERCENT and MSRP_sum_carstate variables. proc summary data=sashelp.cars chartype ; class Make Drivetrain Type Model ; types Make*Drivetrain*Type Make*Drivetrain*Type*Model; output out=step1 sum(MSRP)=COUNT ; run; data want; merge step1(where=(not indexc(_type_,'0'))) step1(where=(indexc(_type_,'0')) drop=Model rename=(COUNT=MSRP_sum_carstate)) ; by Make Drivetrain Type; PERCENT=100*(COUNT/MSRP_sum_carstate); drop _type_ _freq_ ; run; Let's test and see if they make the same values (within 6 decimal places). proc compare data=want compare=final method=absolute criterion=10E-6; run; Results Variables Summary Number of Variables in Common: 7. Number of Variables with Differing Attributes: 2. Listing of Common Variables with Differing Attributes Variable Dataset Type Length Format Label COUNT WORK.WANT Num 8 DOLLAR8. WORK.FINAL Num 8 Frequency Count PERCENT WORK.WANT Num 8 WORK.FINAL Num 8 Percent of Total Frequency Observation Summary Observation Base Compare First Obs 1 1 Last Obs 428 428 Number of Observations in Common: 428. Total Number of Observations Read from WORK.WANT: 428. Total Number of Observations Read from WORK.FINAL: 428. Number of Observations with Some Compared Variables Unequal: 0. Number of Observations with All Compared Variables Equal: 428. Values Comparison Summary Number of Variables Compared with All Observations Equal: 7. Number of Variables Compared with Some Observations Unequal: 0. Total Number of Values which Compare Unequal: 0. Total Number of Values not EXACTLY Equal: 141. Maximum Difference: 1.4211E-14.   ... View more

Hi @js5 ,   There isn't a "return code" or other property to check. What I recommend doing is create a program task that runs at the end of the flow that checks for success (data created/exists, etc.) and then use the script to examine just that one log. ... View more

I think the problem is actually before that.  It looks like when you INPUT the data, it's only grabbing the first 32,767 characters.   I think to make this work, you would need to find an informat to use on the INPUT statement that allows a length >32767.   From a quick read of the docs, I don't see any.   1 data _null_; 2 infile "Q:\junk\input.txt" truncover dsd delimiter='09'x lrecl=100000; 3 length var1 varchar(70000) var2 $10; 4 input var1 var2; 5 6 l=length(var1) ; 7 put l= ; 8 run; NOTE: The infile "Q:\junk\input.txt" is: Filename=Q:\junk\input.txt, RECFM=V,LRECL=100000,File Size (bytes)=60012, Last Modified=14Nov2022:11:41:23, Create Time=14Nov2022:11:41:23 l=32767 NOTE: 1 record was read from the infile "Q:\junk\input.txt". The minimum record length was 60010. The maximum record length was 60010. If you can't find an informat that will read > 32,767 characters, you might have to read it in chunks, as Tom suggested.   ... View more

proc summary data=sashelp.cars noprint nway; var MSRP; class Origin DriveTrain; output out=cars_summary lclm= uclm= mean= /autoname; run; data cars_for_plot; set sashelp.cars cars_summary nobs=nobs; if _n_ =nobs-2 or _n_= nobs-1 or _n_= nobs then ref=1e5; if _n_=nobs then label="LLOD" ; if not missing(MSRP_Mean) then y_datalabel=MSRP_Mean+120000; run; proc sgpanel data=cars_for_plot; panelby DriveTrain / layout=columnlattice noborder noheader columns=3 skipemptycells sort=ascformat; rowaxis type=log label="MSRP"; colaxis label=" " fitpolicy=rotatealways valuesrotate=vertical discreteorder=formatted; keylegend "ssss" / position=right noborder; vbarparm category=Origin response=MSRP_Mean / group=DriveTrain groupdisplay=cluster grouporder=ascending baseline=10 name="ssss"; scatter x=Origin y=y_datalabel/markerchar=MSRP_Mean labelstrip; highlow x=Origin high=MSRP_UCLM low=MSRP_LCLM / group=DriveTrain highcap=serif lowcap=serif; scatter x=Origin y=MSRP / group=DriveTrain jitter markerattrs=(symbol=circlefilled size=5) transparency=0.50 clusterwidth=0.7; refline ref / lineattrs=(pattern=shortdash) label=label labelpos=max ; run;   ... View more

It is even simpler, thank you for pointing me towards highlow: proc summary data=sashelp.cars noprint nway; var msrp; class Origin Type; output out=cars_summary mean= uclm= lclm= /autoname; run; data for_plot; set sashelp.cars cars_summary; run; proc sgpanel data=for_plot; panelby Origin /layout=columnlattice; vbarparm category=Type response=MSRP_Mean / datalabel; scatter x=Type y=MSRP/ jitter; highlow x=Type low=MSRP_LCLM high=MSRP_UCLM; run; ... View more

@js5 :   You are essentially correct.   The "rc=filename(....)" statement is an executable statement - i.e. it won't be honored until the data step begins data processing.  After all, like all functions, it could be a result of an "IF ... THEN rc=filename(...)" statement, which means it can only be executed during actual data processin.   But the INFILE statement in your initial code is effectively a compile-time statement that must be established prior to execution.   Using the "filevar=" option tells SAS that it must be prepared to dynamically assign a physical file to a fileref (even if it is just a single unchanging assignment).         ... View more