|
Nancy,
Ouch! If you intend to have only 395 variables at the end - in other
words, the merging is really an update process, I would suggest doing 48
two-data set merges inside a macro loop. The data sets appear to be fairly
regularly named, so you should be able to do this with out much bother. The
critical data set inside the macro would look something like this:
data sofar;
merge sofar &NEXT;
by program;
run;
HTH - Jack
--
Jack Shoemaker / Oxford Specialty Management / JShoemak@oxhp.com
+ -----Original Message-----
+ From: Patton, Nancy (CAP, RFS) [mailto:Nancy.Patton@gecapital.com]
+ Sent: Thursday, August 27, 1998 12:52 PM
+ To: SAS-L@UGA.CC.UGA.EDU
+ Subject: MERGING too much??
+
+
+ Help SAS-L!!
+
+ I am trying to merge 48 data sets. Each has 395 variables and
+ from 3 to 15
+ observations. There is one common variable: PROGRAM, and that is my BY
+ variable. Here is the error message:
+
+
+ FATAL: Code generation error detected during MISSING smear
+ generation.
+ ERROR: Out of memory.
+
+
+ Here is the code:
+
+ DATA SAVE.SUMMDATA ;
+ MERGE SE1951 SE2951 SE3951 .... SE6981 SE7981 ;
+ BY PROGRAM;
+
+ I am running in batch under MVS, SAS 6.09 TS450. Has anyone
+ ever seen this,
+ and do you know how to get around it??
+
+ Thanks-In-Advance!!
+ -------------------------------------------------
+ * Nancy Patton
+ * (703)553-0737
+ * NKP@asg-inc.com
+ * Nancy.Patton@GECapital.com
+
|
