I just got home and started looking at your code again and a couple of things occurred to me. In particular,      if SEQUENCE_COUNT < LSEQ then output UCST      else output clean.UPDT; might have a problem by not checking for sequence_count > lseq Another thing, you've got IN= set:      merge clean.UPDT (in=a)        SEQ_LST (in=b rename=SEQUENCE_COUNT=LSEQ); but you never test it, so is it possible that records are coming in from seq_lst when there is no corresponding updt? In reference to the memory issue, unless told otherwise, in a data step SAS does it I/O directly - i.e. it reads a record(s) from disk, processes it, and writes to disk. There are exceptions (hash tables for example and Data set options that keep it loaded), but those are exceptions. It helps to remember that SAS was written for mainframes using punch cards (I know this personally). Memory was barely measured in K (hence the reference to K in some of the options), let alone M or G. As I stated before, its PROC SQL that tends to hog memory. Another possibility that updates a data set directly, hence is more dangerous, is the MODIFY statement. http://support.sas.com/documentation/cdl/en/lestmtsref/63323/HTML/default/viewer.htm#n0g9jfr4x5hgsfn17gtma5547lt1.htm Go through a couple of examples and back up you data. ... View more

I'm inclined to agree with KurtBremser, but another possibility is to try UPDATE (SAS(R) 9.2 Language Reference: Dictionary, Fourth Edition). In addition to his suggestions, look at the location of your temporary storage, and clean out any residues directories there. Trouble with the SQL is that it's memory intensive - good chance you'll run into memory problems there., since you mention the 1G limitation, the way to go is DATA step. Other possibilities: have you checked COMPRESS= data set option (SAS(R) 9.2 Language Reference: Dictionary, Fourth Edition) - depending on your data set you might squeeze space out there. Also the length of your variables (SAS(R) 9.2 Language Reference: Dictionary, Fourth Edition). They won't make a difference for temporary data sets (WORK) - SAS expands thing during processing (or maybe that's my 'experience' showing), but will make a difference in your permanent data set. Hope these suggestions help ... View more

I meant to add to check out the ID statement - it lets you retain variables. ... View more