Hi Afor,   When you specify array Growth {*} Jan_2018 -- Dec_2019; to SAS, you mean all the variables in the dictionary of table bls.BLS_PPI2 starting from Jan_2018 to Dec_2019   Example : data foobar;   JAN_2018= 201801; foo=1; bar= 2; Dec_2019= 201912; run; data _null_;   set foobar;   array Growth {*} Jan_2018 -- Dec_2019;   do i=1 to dim(Growth);     put i= growth(i)=;   end; run;   Am i correct if I say that an equivalent of your code without array would be something like this: data BLS_PPI3;    set bls.BLS_PPI2;   format Growth_01-Growth_24 percent10.1;   keep Series_id Growth_:;   Growth_02= (FEB_2018/JAN_2018)-1;   Growth_03= (MAR_2018/FEB_2018)-1;   ...   Growth_24= (DEC_2019/NOV_2019)-1; run;   If yes, variables JAN_2018 FEB_2018 etc DEC_2019 should be in your source table.   hth   Thien ... View more

Hi,   Try Jan_2018 -- Dec_2019 if all these variables are in bls.BLS_PPI2   hth   Thien ... View more

Hi, Very strange screenshot indeed. Took your code and run it with SAS Studio (University Edition) and get for |mydate, mydateNr] = [31DEC9999, 2936547] In SAS, a date is represented by the number of days from January 1rst, 1960. It looks like in your environment, there is a shift of 2 days. You could try this code: data _null_;   d=0; put d= d= date9. d=yymmdds10.;   d=1; put d= d= date9. d=yymmdds10.;   d=2936545; put d= d= date9. d=yymmdds10.;   d=2936546; put d= d= date9. d=yymmdds10.;   d=2936547; put d= d= date9. d=yymmdds10.; run; On my desktop, log shows : d=0 d=01JAN1960 d=1960/01/01 d=1 d=02JAN1960 d=1960/01/02 d=2936545 d=29DEC9999 d=9999/12/29 d=2936546 d=30DEC9999 d=9999/12/30 d=2936547 d=31DEC9999 d=9999/12/31 HTH ... View more

Hi there, Did you notice REAL TIME and CPU TIME in the in the poststed log. NOTE: There were 71085 observations read from the data set DSN.REGISTRY. NOTE: The data set DSN.REGISTRY has 71085 observations and 445 variables. NOTE: DATA statement used (Total process time):       real time           7:25.71       cpu time            2.87 seconds We could interpret REAL time as the total elapsed time for the data step to run, where as CPU time taken by SAS. So where goes the 7:25.71-2.87 difference ? Compared to Doc's log NOTE: There were 71000 observations read from the data set WORK.TEST. NOTE: The data set WORK.TESTCOPY has 71000 observations and 445 variables. NOTE: DATA statement used (Total process time):       real time           0.35 seconds       cpu time            0.35 seconds From the 2 logs, I would probably say that the difference is due to the DSN library vs the WORK library. You could run : LIBNAME DSN list; to see how it is allocated. If it points to a DBMS (ORACLE, TERADATA, MySQL, etc...), the difference is the DBMS. You can try the new SAS 9.2, OPTIONS DBIDIRECTEXEC, maybe it could help. If it points to a network file system, the difference is the bandwidth of the FS. You should then contact your network admin. HTH PS : When dealing with large datasets through SEGuide, I also cleared the default option that opens the result dataset. Save some time doing so. Second thought : Could'nt  be DBMS behind DSN. Won't be able to modfy DSN.REGISTRY this way Ce message a été modifié par : Thien LE TRI ... View more

Hi, If by SAS MC, you mean SAS Management Console, You need to configure a server through SAS MC only if you need SAS functionalities, such as user's authentification through SAS Metadata, etc ... But you can also, for sole file transfer between servers, install any software supported by the 2 two servers, eg winscp, putty, and configure your software as specified. Doing this, the files can be transferred as any binary files, not SAS files (eg datastes, catalogs, etc...) and some features are for you to care Users' authentification file integrity hth ... View more

Hi all, Doc is right.SQL may require multiple steps. And this is to be taken in account when you are dealing with ten of thousands of observations and want to compare all observations from one dataset with all observations from another dataset. This is what we call a cartesian product. "Cartesian Product, you must be out of your mind", some may say. Still, I like the SQL approach. Let's say A is your first dataset and B the second. I would adapt Doc's code like this PROC SQL;  * untested code; CREATE TABLE step1 AS SELECT A.Date_Modified, B.Country,      SUM(           CASE                WHEN  A.Date_Modified BETWEEN B.Event_date-'0:0:30't AND B.Event_date+'0:0:30't then 1                ELSE 0      ) AS Nb_Selected   FROM A, B QUIT; This would give a table like this Date_Modified          Country     Nb_Selected 30/09/2012 09:45:10      UK          1 30/09/2012 09:45:10      US          1 You may want to test this code with a sample, lets'say 1000 obs from each table with the PROC SQL INOBS options : PROC SQL INOBS=1000; If the intermediary result is not necessary, you may want to CREATE VIEW step1 instead of CREATE TABLE step1 Next, you'll have to transpose the intermediary result PROC TRANSPOSE DATA=step1 OUT=result NAME=Country;      VAR Nb_Selected; RUN; to get what you want Date_Modified               UK     US 30/09/2012 09:45:10      1       1 HTH ... View more

Hi, Code and notes do not match Notes say '27JAN2008' that can't be found in code. NOTE: Character values have been converted to numeric values at the places given by: (Line):(Column) 760:16 762:21 762:43 Look in your log, if you see something like 760 if DATE LT "..." THEN It means your DATE is not a character variable HTH ... View more