home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!howland.reston.ans.net!paladin.american.edu!auvm!CORNELLA.BITNET!LMJY
- X-Acknowledge-To: <LMJY@CORNELLA>
- Message-ID: <SAS-L%93010710413719@UGA.CC.UGA.EDU>
- Newsgroups: bit.listserv.sas-l
- Date: Thu, 7 Jan 1993 10:31:28 EST
- Reply-To: Lisa King <LMJY@CORNELLA.BITNET>
- Sender: "SAS(r) Discussion" <SAS-L@UGA.BITNET>
- From: Lisa King <LMJY@CORNELLA.BITNET>
- Subject: Division by Zero in CMS
- Lines: 62
-
- Hi SAS Gurus: I've been off the list for a few months, so please forgive
- me if this problem is old hat. Consider the following simple SAS program
- which I ran under SAS 6.07 under CMS:
-
- DATA;
- INPUT X Y;
- Z=X/Y;
- CARDS;
- 1 1
- 1 0
- 1 1
- ;
-
- And take a look at the SASLOG:
-
- NOTE: SAS initialization used 0.57 seconds and 781K.
- NOTE: 5568K of contiguous memory remained unallocated.
- 1 DATA ONE;
- 2 INPUT X Y;
- 3 Z=X/Y;
- 4 CARDS;
-
- NOTE: The data step has been abnormally terminated.
- NOTE: The data set WORK.ONE has 1 observations and 3 variables.
- NOTE: DATA statement used 0.27 seconds and 512K.
- NOTE: 5108K of contiguous memory remained unallocated.
-
- 8 ;
-
- NOTE: Viobuf storage reached 25 percent full.
- NOTE: The SAS session used 0.88 seconds and 1338K.
- NOTE: 5108K of contiguous memory remained unallocated.
- NOTE: SAS Institute Inc., SAS Campus Drive, Cary, NC USA 27513-2414
-
- Am I crazy, or is this an absurd result? A colleague of mine
- just lost two days of work trying to figure out why his data step was
- "abnormally" terminating. This same program, when run under SAS 5.18,
- produces a note about the value of Z being set to missing due to
- division by zero, but the data set completes.
-
- If any of you can shed some light on this, please help me out.
- I have combed the FM's and all I can find is a passage in the SAS
- Language Guide (v.6 ed.) on page 121 which reads:
-
- "If you try to perform an illegal operation (for example, dividing
- by zero, taking the logarithm of zero, or using an expression
- to produce a number too large to be represented as a floating-
- point number known as overflow), the SAS System prints a warning
- message and assigns a missing value to the result."
-
- Bug? Feature? Is there any way to override this ridiculous problem?
-
- Thanks for any and all suggestions, Lisa
-
- |***************************************************************|
- | Lisa King - Technical Consultant |
- | Cornell Institute for Social and Economic Research (CISER) |
- | Cornell University, Ithaca, New York 14853 |
- | |
- | Bitnet: LMJY@CORNELLA |
- | Internet: LMJY@CORNELLA.CIT.CORNELL.EDU |
- |***************************************************************|
-