I have 30 datasets that are to be sorted and then merged later.I heard there somethng about multi threading where we can sort all 30 data sets at once instead of sorting one after the other.Please help me on this.
what Scott (SBB) is hoping you will find is referred to in many papers as mpConnectfor multi-processing connect. Including this topic the "scalability" focus area, is well worth visiting at http://support.sas.com/rnd/scalability/index.html
/* create TASK1 SAS session to sort the SALES data */
rsubmit task1 wait=no sysrputsync=yes;
proc sort data=sashelp.shoes out=shoes1;
/* get TASK1's WORK library path in parent session */
signon task2 userid=sasprod;
/* create TASK2 SAS session to sort the GOALS data */
rsubmit task2 wait=no sysrputsync=yes;
proc sort data=sashelp.steel out=steel1;
/* get TASK2's WORK library path in parent session */
"getting started with SAS/Connect" needs a little more than the example that you have found.
"signon" works as long as there is somewhere "listening" - so have you a sas server with a service listening for your "signon"?
I suspect not.
You will find many benefits for design and performance of your systems once the SAS/Connect services can be started, but it requires some unusual learning, and probably help from a system administrator for whom setting up system services, like a sas connect listener, is "ordinary".
As well as following Peter's advice I would suggest contacting your SAS Server administrator and/or knowledgeable SAS users at your site. Usually they know what configuration is required to signon. Usually it involves specifying the IP address of the SAS server and what TCP/IP port to communicate on.
BTW the example you have is right on the money. You may also find the WAITFOR statement useful.
The only way of doing concurrent sorting (or any process for that matter) that I am aware of is by doing it in separate SAS sessions. If you have SAS/CONNECT then you can create several asynchronous SAS sessions each doing their own processing in parallel/concurrently then get the parent SAS session to wait until all your child sessions have finished processing.
This is probably only worth doing on a server with the I/O capacity to cope. On a desktop machine the first sort would tend to dominate the disk drive and all of the other concurrent sorts would be very slow.