Karsten M. Self (Karsten.Self@schwab.com) Trilogy Consulting

What part of "gestalt" don't you understand?

WARNING: All e-mail sent to or from this address will be received by the Charles Schwab corporate e-mail system and is subject to archival and review by someone other than the recipient.

> ---------- > From: Self, Karsten > Sent: Tuesday, November 10, 1998 11:24 AM > To: 'Ashiru, Babatunde'; 'SAS-L' > Subject: RE: Funky Stuff in SAS? > > A 12 minute sort bombing like yours did leans pretty heavily toward > banging into the 2 GB filesize limit. > > Run the following korn shell commands after finding your work library > (assumed to be $saswork): > > while : > do > ls -lt $saswork > sleep 15 > done > log.dat > > (to get to korn shell if you're not there alread, execute /bin/ksh on > the command prompt. HPUX uses a korn-like shell by default, if your > SHELL environment variable is /bin/sh or /bin/ksh, you'll be fine). > > ...which will list the files in your SAS work library to a logfile at 15 > second intervals. You'll have to kill the logger manually when your SAS > job bombs out. I'm betting you'll find a file approaching 2 GB. The > error messages are consistant with what I've experienced, usually > followed by "file is damaged" (paraphrased) and some other stuff. > > To get around this, use SAS large file support. You nead to specify a > couple of options to load the module and specify a partitioned WORK > library, it's pretty painless. Look under 6.11 changes and enhancements > in online help or call SAS tech support to set this up. There've been a > couple of SUGI papers on the subject in the last few years as well. > > > Karsten M. Self (Karsten.Self@schwab.com) > Trilogy Consulting > > What part of "gestalt" don't you understand? > > WARNING: All e-mail sent to or from this address will be received by > the Charles Schwab corporate e-mail system and is subject to archival > and review by someone other than the recipient. > > > ---------- > > From: Ashiru, Babatunde[SMTP:AshiruB@whiteoaksemi.com] > > Sent: Tuesday, November 10, 1998 11:13 AM > > To: 'Self, Karsten'; SAS-L (E-mail) > > Subject: RE: Funky Stuff in SAS? > > > > Good question, Karsten. I run SAS 6.12 in a UNIX box on tired HP > > server ... > > but as I mentioned in an earlier posting, in reply to Bernard Tremblay > > (sp), > > we moved the process that generates this error message to another > > server > > that's less swamped and still status quo on the I/O problems. > > > > > > -----Original Message----- > > From: Self, Karsten [SMTP:Karsten.Self@Schwab.COM] > > Sent: Tuesday, November 10, 1998 2:07 PM > > To: SAS-L@UGA.CC.UGA.EDU; 'Ashiru, Babatunde' > > Subject: RE: Funky Stuff in SAS? > > > > Set OPTIONS NOVP to get rid of the funky triplicate error > > messages. > > This is intended for old-style lineprinters which could > > overprint a > > line > > to embolden it. Doesn't do much for a laserprinter these days. > > > > This is similar to the messages generated when you push a file > > over > > the > > 2 GB filesize limit on 32 bit Unix. You don't mention your OS, > > and > > I > > haven't had the experience yet under Windows, but you might > > check > > the > > total size of your file and/or temporary sort files. Sort > > typically > > multiplies total file size by 2.5, though I haven't checked to > > see > > how > > big the individual sort files become. > > > > What's your OS? > > > > Karsten M. Self (Karsten.Self@schwab.com) > > Trilogy Consulting > > > > What part of "gestalt" don't you understand? > > > > WARNING: All e-mail sent to or from this address will be > > received > > by > > the Charles Schwab corporate e-mail system and is subject to > > archival > > and review by someone other than the recipient. > > > > > ---------- > > > From: Ashiru, Babatunde[SMTP:AshiruB@WHITEOAKSEMI.COM] > > > Reply To: Ashiru, Babatunde > > > Sent: Tuesday, November 10, 1998 10:55 AM > > > To: SAS-L@UGA.CC.UGA.EDU > > > Subject: Funky Stuff in SAS? > > > > > > Hi, all, > > > > > > SAS is doing some funky stuff or is this unique to me? > > > First of all, I am still getting the following error message > > when > > SAS > > > tries > > > to Sort a dataset of about 420,000 plus records: > > > > > > "ERROR: An I/O error has occurred on file > > PRDCTNDS.CHIP_NW.DATA.ERROR: > > > An > > > I/O error has occurred on file PRDCTNDS.CHIP_NW.DATA.ERROR: An > > I/O > > > error has > > > occurred on file PRDCTNDS.CHIP_NW.DATA. > > > ERROR: An I/O error has occurred on file > > PRDCTNDS.CHIP_NW.DATA.ERROR: > > > An I/O > > > error has occurred on file PRDCTNDS.CHIP_NW.DATA.ERROR: An I/O > > error > > > has > > > occurred on file PRDCTNDS.CHIP_NW.DATA. > > > NOTE: The SAS System stopped processing this step because of > > errors. > > > NOTE: SAS set option OBS=0 and will continue to check > > statements. > > This > > > may > > > cause NOTE: No observations in data set. > > > > > > WARNING: The data set PRDCTNDS.CHIP_NW may be incomplete. > > When > > this > > > step > > > was stopped there were 0 observations and 29 > > > variables. > > > NOTE: PROCEDURE SORT used: > > > real time 12:50.78 > > > cpu time 33.58 seconds > > > > > > > > > > > > 127 data pubsasds.frlchip(compress=yes); > > > 128 set prdctnds.chip_nw; > > > 129 > > > > > > NOTE: The data set PUBSASDS.FRLCHIP has 0 observations and 29 > > > variables. > > > WARNING: Data set PUBSASDS.FRLCHIP not replaced because of > > NOREPLACE > > > option. > > > NOTE: DATA statement used: > > > real time 1.73 seconds > > > cpu time 0.48 seconds" > > > > > > What happens next is a complete mystery to me: Upon trying to > > append > > > new > > > records to the supposedly large dataset, SAS throws away the > > huge > > > dataset > > > and starts the new master dataset with the proceeds of the > > transaction > > > dataset, i.e., if the master dataset has something like > > 420,000 > > > records in > > > it and the new transaction dataset has 406 records, the append > > > statement > > > should cause master dataset to grow by 406 records; however, I > > am > > > completely > > > befuddled as to why sas abruptly replaces the master dataset > > with > > new > > > records from the transaction. This is very disturbing. I > > have > > > observed > > > similar incident happened before but thought it's nothing but > > a > > fluke > > > or > > > perhaps it's an artifact of something completely obsucred but > > I > > > literally > > > watched as it happens again before my eyes ... WHAT's going > > on, > > guys? > > > > > > > > > Any ideas on how to resolve this mystery will be highly > > appreciated. > > > TIA > > > > > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > > ++ > > > Babatunde R. Ashiru > > > White Oak Semiconductor, Product Sustaining Group, > > > Analysis > > > Tools Section > > > 6000 Technology Boulevard, Sandston, Virginia 23150 > > > Phone: (804) 952-7944 Fax : (804) 952-7902 > > > EMail : <mailto:ashirub@whiteoaksemi.com> > > ashirub@whiteoaksemi.com > > > Pager: 1-800-759-8888, PIN 1589371 > > > > > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > > ++ > > > "Mental geography is not linear, and cannot be mapped onto a > > flat > > > surface without severe distortion."--Larry Wall, Creator of > > Perl > > > > > >