Would something like below do the job for you? data have; length my_key_var 8; array vars_ {200} 8; my_key_var=1; run; proc sql; create table abc like have; quit; proc datasets lib=work nolist nowarn; modify abc; ic create my_key_var_pk=primary key(my_key_var); run; quit; proc append data=have base=abc; run;quit; proc contents data=work.abc; run;quit; ... View more

At a second glance, your answers add nothing, you're supposed to explain yourself here, not send people to other pages that they can find with Google if they want to.   Nevermind, if you don't want to or most likely can't provide a valuable answer, then don't. What you posted is wrong, it's not a solution to move a file from Unix to Windows using SFTP, so your answer is just misleading Googlers like me who end up here.  Even if your code is right, it's not easy to adapt it for the case of the users who come to this page for help because you provided the least possible amount of explanation for us to make sense it.   No wonder your answers got 0 likes. ... View more

Sorry, I misunderstood your request, so ignore my previous reply. To do what you want is a special case different from displaying the segment value using SEGLABEL. Your trying to show the cumulative values in the bar. To do this, you will need to do some data processing outside of SGPLOT. Here are the basic steps:   1. Transpose your data as described by @Jagadishkatam , but using the X1 and Y1 variables. 2. Pre-summarize your data using PROC SUMMARY or PROC MEANS.setting CLASS to be DATE and NEWC. 3. Using that data, calculate the middle value in each segment (we'll call that var "middle"). At the same time, accumulate the sums for each bar segment and put it in a variable called "seglabel". 4. Render using SGPLOT with something like this:   proc sgplot data=whatever; vbarparm category=date response=sumvar / group=newc; text x=date y=middle text=seglabel; run;   Hope this helps! Dan ... View more

This does what you want though I am unsure why you'd want the 2018 entries to take the 2017 data. proc sql; select a.ID, b.DATE, Y.BALANCE, Y.BAL from (select unique ID from X) a cross join (select unique DATE from X) b left join Y on Y.ID=a.ID and intck('qtr.3',Y.DATE,b.DATE) in(0,4) order 2,1; quit; ID DATE BALANCE BAL 1 01/12/2017 200 A 2 01/12/2017 500 B 3 01/12/2017 700 C 4 01/12/2017 800 C 5 01/12/2017 .   6 01/12/2017 .   7 01/12/2017 .   1 01/01/2018 200 A 2 01/01/2018 500 B 3 01/01/2018 700 C 4 01/01/2018 800 C 5 01/01/2018 .   6 01/01/2018 .   7 01/01/2018 .   1 01/02/2018 200 A 2 01/02/2018 500 B 3 01/02/2018 700 C 4 01/02/2018 800 C 5 01/02/2018 .   6 01/02/2018 .   7 01/02/2018 .   1 01/03/2018 200 A 2 01/03/2018 500 B 3 01/03/2018 .   4 01/03/2018 .   5 01/03/2018 700 C 6 01/03/2018 800 C 7 01/03/2018 .     ... View more

Since you can't use %include (since embedded programs -- no file for server to reference), rather than referencing Program2 in Program1, you could draw an explicit link (right-click Program2 and select Link to... or hover mouse on right side of Program2 node, click and drag link line to Program1) from Program2 to Program1.  For example: Then, whenever you run the project, process flow, or branch, Program2 will always run prior to Program1.   Casey ... View more

Thank you. Its resolved. It had length mismatch between the values in one of the var in sas and var in db2 table. ... View more

They both are the same code. I could expect the same performance either . ... View more

Thanks. My SAS server is grid server. I came to know that I can not import the access db ditectly to SAS GRID server. I am able import it in PC/BASE SAS using proc import and libname statement.   ... View more

This seems like a lot for your purpose but I'm sure it'll work.  If it's a large dataset I'd put in a where clause too.  Scan() will separate by the blank you want to keep and catx() will strip each variable but use the space as a deliminator. Hope it helps: data have; infile cards dsd; input char $; cards; abc 25 spc 26 xy 20 ; run; data want; set have; cchar = scan(char,1,''); nchar = scan(char,2,''); if _N_ => 1 then do; new_var = catx(' ',cchar,nchar); end; run; ... View more