Date:         Thu, 20 Sep 2007 19:13:50 -0000
Reply-To:     kvasikonkav <kvasikonkav@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         kvasikonkav <kvasikonkav@GMAIL.COM>
Organization: http://groups.google.com
Subject:      Re: Large datastep interrupted,
Comments: To: sas-l@uga.edu
In-Reply-To:  <8AD8F86B3312F24CB432CEDDA71889F203EF7DA6@ex06.GHCMASTER.GHC.ORG>
Content-Type: text/plain; charset="us-ascii"

On Sep 20, 11:09 am, parde...@GHC.ORG ("Pardee, Roy") wrote:
> But even if sas will use it w/out complaining, it is going to be
> suspect, isn't it?
>
> -----Original Message-----
> From: SAS(r) Discussion [mailto:SA...@LISTSERV.UGA.EDU] On Behalf Of
>
> Gerhard Hellriegel
> Sent: Thursday, September 20, 2007 12:44 AM
> To: SA...@LISTSERV.UGA.EDU
> Subject: Re: Large datastep interrupted, have a large data1.sas7bdat.lck
> file - any way to convert to usable data?
>
> Try to rename it by removing the .lck and have a look into the dir
> window of the library. Maybe the dataset is usable. If not, try to
> repair it with proc datasets.
> Gerhard
>
> On Wed, 19 Sep 2007 23:04:03 -0000, kvasikon...@GMAIL.COM wrote:
>
> >I was running a job on a server, and the server was rebooted.
>
> >Consequently, I have a partially completed datafile, which has  the
> >suffix .sas7bdat.lck
>
> >The ".lck" file is a 225,000,000kb file (I think that is 225gb,
> >right?)
>
> >SAS was nearly done processing a datastep when the server was rebooted.
>
> >My understanding is that the ".lck" extension is sort of a temporary
> >dataset that will later be renamed once the datastep is fully finished.
>
> >I don't want to wait another 4 days for the program to run again, so I
> >was hoping the nearly complete ".lck" data could be recovered into a
> >regular dataset without too much effort?
>
> >is there any way to recover any of the data out of a ".sas7bdat.lck"
> >file, or convert it into a sas7bdat file?
>
> >Thanks!
>
> >Thanks again!

Thanks everyone for the suggestions.
I followed the advice, and ran it in 10 separate datasteps.
Unfortunately, the program is now creating I/O errors, so i suspect
that something is fundamentally wrong (aside from the server beign
spotty).

Thanks again for the responses.