04-11-2016 03:27 AM
I received a SAS dataset created using SAS Enterprise Guide from the client. I need to access it on my company's SAS server using SAS 9.2. However I am getting an error while accessing the same:
ERROR: File Filename.DATA not compatible with this SAS version.
Has anyone faced the same issue? How do I resolve it?
04-11-2016 03:41 AM
Most likely the SAS dataset was either created under a different operating system to yours or a different "bitness" (32 bit versus 64 bit).
One method to transport datasets affected this way is use the CPORT procedure to create a transport dataset on the originating SAS system, then use the CIMPORT procedure to read it in on the target SAS system. Check the documentation on these procedures.
04-11-2016 03:54 AM
Thanks for your suggestion. However, that would require me to go back to the client and request him to run the code again and resend it. Is there no other way to deal with it on our SAS 9.2 server itself?
04-11-2016 04:55 PM
If the dataset came from a different operating system / bitness then it can only be fixed by exporting from that OS. If your dataset source is the same operating system then the issue raised by @KurtBremser could be another cause.
04-11-2016 04:11 AM
This: http://support.sas.com/kb/49/496.html might describe the reason for your problem.
In this note, SAS Support only offers the same two remedies: recreate the data set or use PROC CPORT/CIMPORT.
I personally don't think it was a wise decision by SAS to have that option turned on by default, when it renders datasets incompatible with previous versions.
My to-do list for the migration to 9.4 now has a new item: disable EXTENDEDOBSCOUNTER.
04-11-2016 02:37 PM
I think this is the most likely issue. Can you find out if your client is running a later release than you are?
In each release, SAS has added a few new features to data sets that, if used, make the data set incompatible with earlier releases. Here is a list of these features: http://support.sas.com/rnd/migration/planning/files/forward.html
Links on this page go to a more detailed page with available workarounds.