proc ginside work files

[RealePrimavera]

Dears,

Any experience with proc ginside errors? 

I am receiving this even after configuring my utility files from work library to local space:

"ERROR: Insufficient space in file WORK._GINSID000000000000000000004.DATA.
ERROR: File WORK._GINSID000000000000000000004.DATA is damaged. I/O processing did not complete."

Earlier I changed utility path for intermediate files from work to local disk to overcome this space problem.

Most confusing thing is for the other processes I can see utility files being created in diverted local path; however, with ginside this error pertains. 

Not sure, how proc ginside manages to create intermediate files in the work lib. 

Code is very basic:

proc ginside data=have /*file with latitude longitude information*/

                     map=shapefile

                     out=want;

id regionlabel;

run;

[Reeza]

If you have a full licensed version of SAS I would suggest working with SAS Tech support on this. GINSIDE is a resource intense process, so do it in chunks may also speed things up for yourself. Have you tried doing only a few points at a time and getting any results?

[RealePrimavera]

Thank you @Reeza !!!

[Kurt_Bremser · Accepted solution]

As you can see by the .DATA extension, this is not a utility file, but a temporary SAS dataset, and that is why GINSIDE puts it into WORK. You will have to make place for this dataset in WORK if you want to run the procedure for all your data in one swoop.

Run it with a subset and see if it works. Record consumption in WORK during this. Run it with a larger subset, and again record storage consumption. You can now make an educated guess how much space you will need. Check if that is available on the WORK volume, and check if other limits (quotas, user limits) are present.

[TomKari]

In addition to all of the above suggestions, which make great sense, if you have full control over your SAS environment you can relocate your WORK library similar to how you moved your UTILLOC.

Tom

[RealePrimavera]

Thank you @TomKari ! 

That's what my next planned step is, after @Kurt_Bremser clarified that those are temp. data-sets. 

All your replies helped a lot to find solutions quicker.