08-01-2017 12:21 AM
08-01-2017 01:02 AM
If yor code is one step only you can use libname USER assigned by yourself to a local big space, like:
libname user '...disk area under myfolders ...';
no need to change your code. sas will use libname USER automatically instead WORK.
If there are many steps you can add in between steps of deleting non used datasets, for eample:
data c; set a b; run; /* no more use of datasets a b */ proc datasets lib=work nolist; delete a b; run; quit;
08-01-2017 01:40 AM
08-01-2017 02:24 AM
@Reeza, assuming this is the case, should next code - being first code to run - help?
proc datasets lib=work kill; run; quit;
No, that only cleans the current WORK directory. Orphaned WORK directories from previously crashed sessions must be cleaned with the cleanutil utility, but that can't be used on SAS UE (no XCMD or console access). That's why I posted my Ballot Idea that UE cleans its temporary spaces on its own when it is (re)started.
08-01-2017 02:18 AM
You can add your vote on https://communities.sas.com/t5/SASware-Ballot-Ideas/Do-a-thorough-clean-in-SAS-University-Edition/id..., which deals with this kind of problems.
08-01-2017 02:31 AM
SAS UE is a learning edition so you should be able limit your data sets drastically and still learn.
Yeah, but 491547 observations (as stated in the OP) should be far removed from the 10 GB space limit of SAS UE.
If the OP kept lots of intermediate datasets in WORK up to this point, inspecting and cleaning the WORK may help. But if it's the problem with the leftover WORKS, a reinstallation of SAS UE might be in order.