Hi, Take a look at this paper -- it shows PROC TABULATE with BY group processing and PROC UNIVARIATE with BY group processing. The goal in this case was to organize all the BY group output together in the new output (instead of by procedure). Look at https://support.sas.com/resources/papers/sgf09/318-2009.pdf on page 7 at Figure 11 to see the default way the output from two procedures is organized with BY group processing, compared to page 11, Figure 21 to see the rearranged output. Where you have SGPLOT and TABULATE in your example, I have TABULATE and UNIVARIATE. But the underlying concept is the same. I still think the macro approach is simpler.     I haven't looked at the link @Reeza gave you, but I suspect it is something similar to what I have below: %macro getzip; OPTIONS NODATE NONUMBER; ODS GRAPHICS / RESET=ALL; ** get list of unique zip codes; proc sql noprint; select distinct(zip) into :wantzip separated by '~' from monthly_sales order by zip; quit; %put &=wantzip; **count number of separators and calculate number of unique zips; %let cntsep = %sysfunc(count(&wantzip,~)); %let cntzip = %eval(&cntsep+1); ** start do loop to cycle thru list of zips; ** and create the zip code macro var for each loop; %do i = 1 %to &cntzip; %let zip = %scan(&wantzip,&i,~); %put This is for by group where zip = &zip; ** use zip code macro variable for file name and WHERE and TITLE; ODS PDF FILE="C:\TEMP\TWOTHINGS_&zip..PDF" STARTPAGE=NO STYLE=EGDEFAULT nogtitle; ODS RTF FILE="C:\TEMP\TWOTHINGS_&zip..RTF" STARTPAGE=NO STYLE=EGDEFAULT nogtitle; ODS HTML PATH="C:\TEMP" (URL=NONE) FILE="TWOTHINGS_&zip..HTML" nogtitle OPTIONS(PAGEBREAK='NO') STYLE=EGDEFAULT; title "Monthly Sales for Zip Code: &zip"; PROC SGPLOT DATA=Monthly_sales; where zip = "&zip"; VBAR cty/GROUP=type GROUPDISPLAY=CLUSTER; XAXIS TYPE=DISCRETE; RUN; TITLE 'PROC TABULATE'; PROC TABULATE DATA=Monthly_sales; where zip = "&zip"; CLASS cty type; VAR sales; TABLE cty ALL, type='SALES STATISTICS BY TYPE'*sales=' '*(N MEAN); RUN; TITLE; ODS _ALL_ CLOSE; %end; %mend getzip; ** invoke macro program; %getzip; First, I make a package of program statements, called a macro program definition. The name of my macro program is %GETZIP. When the macro program is executed, the first thing it will do is make a macro variable called &WANTZIP from the monthly_sales data file. For your sample data, &WANTZIP will look like this: WANTZIP=52199~52200~52388~52423     Then the %DO loop will cycle through the list and extract 1 zip at a time into the macro variable &zip -- that macro variable is used primarily in the FILE= option, the TITLE and the WHERE statements. The ODS statements are INSIDE the macro program and so, using &zip in the FILE= option gives you a separate file for every unique zip code.     In my opinion, this approach or a macro program with a CALL EXECUTE in a DATA step is an easier approach than using ODS DOCUMENT, since you already had a working SAS program to start from.   Cynthia Cynthia ... View more

Hi: I think you mean this previous posting: https://communities.sas.com/t5/ODS-and-Base-Reporting/Mail-merge-randomization-forms/td-p/172172 Cynthia ... View more

It helps to show the desired result. Since your desired result is date of execution dependent it would be a very good idea to explicitly state the date of the program execution for future reference when a search leads someone to this thread.   I have to say I am not quite understanding what all your April references are getting to. I suppose it has something to do with why 05Jan2018, 28Feb2018 and 09Feb2017 is excluded but the reasoning is not quite clear. ... View more

Sorry my post did not make much sense I just realized. Here is what my data situation and what I want. Thanks for your help. My fiscal year goes from March 1 to end of February. At any given day (let's say today) I have my current fiscal year data until today (Mar 01 2018 - May 19 2018) and data from last 2 full fiscal year. For example today I have: FiscalYear 2018-2019: Mar 01 2018 - May 19 2019 Fiscal Year 2017-2018: Mar 01 2017 - Feb 28 2018 Fiscal Year 2016-2017: Mar 01 2016 - Feb 28 2017 WHAT I WANT: ------------------- Filter the data until the end of last month (in my case April of 2018-2019 Fiscal Year) and I want the code to determine that from system date. Filter the data from each last 2 fiscal year until the end of April of each year. In my example it will be April 2017 & April 2016. Every month when I will run my code, I don't want to change my Year/Fiscal year manually on my code. My output table should look like this: 16Mar2016 Bio2 01Mar2018 Eng1 15Mar2018 Bio1 27Apr2018 Bio1 HERE IS MY Example DATA DATA have; INPUT Date Course_ID $4.; INFORMAT Date Date9.; FORMAT Date Date9.; DATALINES; 05Jan2018 ENG1 16Aug2017 Bio2 11MAY2015 Eng2 28Feb2018 Geo2 16Mar2016 Bio2 01Mar2018 Eng1 09Feb2017 Che1 19Jan2015 Phy2 15Mar2018 Bio1 12Mar2015 Phy2 19May2018 Che1 27Apr2018 Bio1 ; PROC PRINT NOOBS; RUN; ... View more

If your fiscal year began in January, you could just use the YEAR function to get fiscal year, as in      fiscal_year=year(date1);   But since your fiscal year begins in March you could do the following: (1) get the closest march preceding DATE1, and (2) get the calendar year of that march, and call it fiscal year, as in:   data want; set year17_18 year18_19; fyear=year(intnx('year.3',date1,0,'B')); run;   The INTNX functions says to add a some units called "year.3"  (years beginning in March) to DATE1.  The 0 as third argument says to add zero such year, and the 'B' say to aligned the resulting date to the beginning of the year.3 time span.  So Dec20, 2017 --> March 1, 2017  --> FYEAR=2017.   I've labeled fiscal years using the calendar year of their first month.  So fiscal year March 2017-Feb 2018 is 2017. ... View more

i @mlogan   Not sure if you resolved the issue, if not:   1st, please have a look at Usage Note 18363 (Troubleshooting guidelines for successfully sending an SMTP e-mail from SAS® software). Can you send an email that way.   Then using instructions and sample code from Usage Note 19767 (Using SAS® software to send SMTP email), does it work?   Hope that helps. Cheers, Damo   ... View more

Thanks Tom. I did figure that out. Yes, you are right ATTRIB and LENGTH really worked well at the end. ... View more

What output destination are you using and how are you reporting your data?   ie XLSX via ODS EXCEL using PROC REPORT ie PDF via ODS PDF using PROC TABULATE     ... View more

@mlogan wrote: Hi art297, Your code just copy the column STATE to STATE2.  No it doesn't.    The code is: data want; set have; retain state2; if not missing(state) then state2=state; run;   RETAIN holds the value until it is changed explicitly. The assignment of State to STATE2 only occurs when there is a new value of STATE.  It returns exactly your desired stated output using the simplest methods.   If it isn't doing what you want, you need to provide more details.   ... View more

Try the informat of dollar9. The informat should reflect what the variable currently looks like and apply a format to show how you want it displayed. ... View more