As you have an official version you can call TS for that. There where some notes with updates to w8 did you have seen those? ... View more

Every libname and every sql will open one or more sessions threads by default. The sharing of sessions will limit the number. When using td volatile tables you will need a full shared one (global) Not needing those libraries in all cases then use defer=yes  it will open only when needed. Also handy on delaying the autentication. Slow long running tasks is an other possible cause running for days ublimited by not having the correct options like bulk load will not release resources. ... View more

The sas session is just the program that is running to process your SAS code. As you could start and use several of those running parallel do not think as a single approach. ... View more

RafaelZ, please forget SAS/ACCESS ODBC as in improvement in contrary as you miss all teradata options you have availabel now you can expect more issues. The mode teradata / ansi (TD environment) is a basic option you should understand as the same with the SAS/ACCESS tD specifics.  You cannot circumvent that need for knowledge switching tools.  http://support.sas.com/documentation/cdl/en/acreldb/67589/HTML/default/viewer.htm#n15q85z7nwkl3pn189ekttbnnzbe.htm ... View more

Chris, There are major differences between threads and forks. It something fundamental for system concepts. A nice descritpion of that is at http://www.geekride.com/fork-forking-vs-threading-thread-linux-kernel/  . The forking is the one the spawner does, that is creating new visisble sas sessions. The OS session manager (workload balancing) will see sessions as the major anchors points (one kind of  table) wihtin each sessions threads being run.  By the sharing nature of threads that is less likely being well possible.  http://www.thegeekstuff.com/2011/04/ps-command-examples/  (see the pid/tid difference at example 6) Yes you can see some effects as the more spread used of cpu's instead a very skewed (single) usage. In a multi-users environment with many users active and having many cpu's that is different than a desktops single user approach.  For fun the OS geek is assuming the application bein the SAS-system.   SAS applications however are built on the SAS-system. The confusion is that SAS developers are not seeing that much of the OS-level like threading/forking. However designing efficient code they shoud understand the basics of those.   ... View more

When it is just for calculating some value  out of some other variables that there are many options. 1- build a macro library for including needed code modifications.     When going macrotized that can be build to an very advanced level 2- use the %include statement  -  %inc fileref(menber) ;  - for bringing in building blocks of code.    Doing this the code blocks must be very standardized and simple. The technical solution is that simple 3- use FCMP or other precompiled function options.      ... View more

You are using a dedicated type of raw data. that dsd.  That one behaves with ignoring the length specifications of the given formats. When lengths (char) a needed you must specify those. Would you use raw data in a classic column oriented way that the lengths of formats are not ignored. The order of variables is a weird thing where all dependicies on the logical order should be avoided. the physical order has been changed with v8 I believe by combining the numerics and chars on word boundary where it was on bytes. The reason is that the technical architecture of CPU and IO has drastically changed by time.  The logical order of sas variables is concept of the 70's. It says a variable gets defined as soon the language compiler detect it. By that ordering of variables is influenced by coding.  ... View more

the default mode for TD SQL assistant is MODE=Teradata , the default with SAS/ACCESS is MODE=Ansi. Switching the mode is not only a matter of syntax but also of behavior.  Every SQL statement with TD should be followed by a commit in Ansi mode to make the updates persistent. That is not needed with mode=Teradata. You could see that at the examples using the execute syntax using explicit pass through. Getting all the data to SAS with a select * is a bad idea as TD is parallel processing all data where SAS doesn't.  Using volatile tables in TD will require a global SAS/TD session. The combination TD/SAS is nice but is requiring some skills for the dedicated behavior. ... View more

This is possible with external DBMS systems as being a job of the DBA. One common approach with that is the creation of views. With SAS you could have similar views build with information map studio for restricted programmed users access (Stored processes). The bottom line is that none of them are structural approaches that can become easily a part of some business process.  The approaches with SAS are more cosmetic as fundamental.    A better approach could be splitting the table according security requirements. and only join the tables (may be a view) when needed.  ... View more

Using parallel processing is one of the many techniques to improve turn around times for processes. As a rule of thumb do that in this order of precedence: 1/ look at your used algorithm. Doing anything that is not necessary needed is waste/     Understand the possible bottlenecks. 2/ choose of possible available solutions like already programmed ones the one that is most  appropriate.     The difficulty is getting to know the ones you do not know without losing too much time. 3/ re-evaluate sas options.  It should be part of the installation/configuration but there can be missed a lot. 4/ re-evalutate operating system and hardware options it should be part of the design and set up but also there a lot could have gone wrong? With proc means you are IO-bound as that is the bottleneck you cannot go faster as one pass for all. The cpucount / threading will not help at SAS only the operating system will do some thing as IO can be done delayed/predicted (caching). Base SAS(R) 9.4 Procedures Guide, Third Edition (proc means threading on input data). When you would have many smaller datasets and the weighted options is functionally acceptable. you could imaging to pre-calculated on all those datasets generating a weighted variable with intermediate results. and then combine those results (less data) use the weight. That is similar to a map/reduce approach. ... View more

Ronan, there is no need to do something special on Windows home dirs with roaming profiles as those issues are well understood by windows system admins. Even the Saswork to a local drive location is no problem as that one is also known as a generic one. The problem arises when going server based or going Unix based. The problem exists because is not clear in their requirements with user sessions and user profiles. That is a root cause of a lot. Sas is also not following the  common Guidelines to work on Linux Unix systems just see those and see what Sas is doing. The ibm paper made with Sas is clear on the omission on setting ulimits. Knowing the system level it is possible to correct some things. Ulimits can be set at the usermods.sh script. Sas made a note on that. The /home filesystem is a common issue Unix os admins are complaining on. It has the storage for system and all service and all user accounts. By that a user could block the system or service accounts using all space. That problem is eliminated by setting up different file systems for each class of accounts. The need or question for quotas is likely eliminated with that. Cgroups is a total different management topic as it is about workload balancing of resources like memory and cpu. The goal is sharing many different user groups with a minimum agree service level. Allowing many possible resources to be used is optimizing the harware.  Why being a dictator disabling functionality of being afraid no capable managing the users and resources? ... View more

Why limiting users as you can isolate them by groups and filesystem? That are common approaches on the s level. Put all users in a group below home in a separate filesystem. Have their code in separate filesystem and their data. Their you have you quotas. Within webdav and sas metadata there is no seperation possible. As you need to segregate in dtap and regular users vs analysts you could need many of those. The machines goals hold be delivering work an service not by limiting and hurting the users. ... View more