On a project we have focused on delivering content via Web Report Studio and Infomaps.

We now want to allow users to access the content via Office Addin and Enterprise Guide.

But (there is always a butt we only want users to access data via Information Map, we don’t want them to access the base libnames or tables.

Why you ask, because we have all the business rules embedded in the Information Maps so we don’t want users bypassing these and defining there own business rules on the base data.

Of course if we deny access to the libname then the Infomaps will fail. We can’t restrict access to all data types (i.e tables) in AMO or EG.

So any ideas out there?

Things we are going to try:

• Implement workspace server pooling (grant access to tables trusted user, but not actual use)
• Create a workspace server for WRS reports with full rights inherited and a workspace server for AMO/EG users with linbame rights restricted

But we are pretty sure that neither of these will work.

As the title says, Help!

What this says is that due to the libname bing a metadata based libname you can’t write to it.

Well there is an easy way to fix this:

1. In SAS Management console go to the properties of the desired libname
2. Click on the extended attributes tab
3. Create an new entry with the following values
   1. Name = AssignMode
   2. Value = 2
4. Try it via EG now and Helens your auntie (but not your prime minister – apologies to non kiwis its a GIS joke!)

This lovely option is easily found here:

http://support.sas.com/kb/20/760.html

Other option values you can use are:

0 Assign using SAS Enterprise Guide
1 Assign using the META engine, METAOUT=ALL (default META engine behavior)
2 Assign using the META engine, METAOUT=DATA
3 Assigned by the SAS server (pre-assigned)
This info (and a raft of other cool Enterprise Guide stuff) can be found here:

http://support.sas.com/techsup/technote/ts788.pdf

I can understand why you would install and run SAS Enterprise Guide, but haven’t quite seen the value in Vista, thank goodness Dell now ships there PC’s with Windows XP.

Now if only SAS would release Enterprise Guide for my Mac (5% market share and growing

One of the tips has ended up saving me a lot of time lately.

When you are in Enterprise Guide and you have large flows in the project designer, you also typically end up with a large number of tables, result, logs, code blocks etc open as well as individual tabs.

The downside of this is it is a hassle scrolling left to go back to the procces designer view.

I thought I was being clever when I worked out I could just click on the project explorer view and then on project designer at the top of the tree.

Well all you need to do is press ‘F4′ and you will be taken to the project designer windows faster than a F1 Ferrari.

When a user configures their connection to the SAS Server in AMO they can save their password, so they effectively gain a form of single sign on. (The password is stored as an encrypted text string in an XML file).

A number of customers I talked to also have some form of LDAP authentication setup (i.e. Active Directory), Unfortunately when a user changes their password on the LDAP server, AMO doesn’t know about it. It keeps trying to authenticate the user with their old password until the users account gets locked.

SAS Enterprise Guide also enables the user to store their connection credentials, but it seems to prompt the user to re-enter their credentials if the authentication with the server fails, therefore the users account doesn’t get locked.

We are working through some work arounds for this to see if we can fix the AMO issue, but has anybody else struck this?

Anybody else fixed it?

Its often said it is ok to make a mistake, but not to make it twice as you haven’t learnt from your mistakes, well here is one mistake that I constantly make when doing the above process (I blame my failing memory due to age).

Enterprise Guide allows you to create new datasets where the variable (column) name has a space in it. The SAS Metadata Server does not like these at all. The problem is it imports the metadata for the dataset fine, its just when you go to view the data all you see is the number of rows, but no columns and no data.

It doesn’t matter if you access the dataset in Enterprise Guide ( via the Metadata Server of course) or use view data in the SAS Management Console, all you see is blank rows.

But drag the exact same table directly into Enterprise Guide and its all hunky dory.

Of course I then spend ages going through metadata security assuming I have read metadata rights but not read data rights and that is whats causing the problem.

Finally I will work out that I had a space in the column name, fix it and wallah. So I thought i’d blog about it just in case anybody else strikes the problem (or I at least remember to check my blog next time).

Oh and if you try and view data in Data Integration Studio you will get this error:

Warning: No rows returned by columns query for table DIM_ORG , no columns registered.
Warning: No rows returned by columns query for table DIM_ORG_OPERATIONS , no columns registered.

(at least one of the SAS interfaces reports that their is actually a problem, yah DI development team!)

I asked around for the best option and was told that using ‘_n_’in a datastep would do i, but of course this would have to be a custom code node, and I was using the Enterprise Guide filter and query wizard, which meant I would have to create a second step within my process flow.

No biggy, but within Oracle I would have just called the Sequence function and all would have been good,

Then our resident SAS Geek (note the removal of Uber) mentioned that I could use ‘Monotonic()’. So I created a new computed column in the query, added Monotonic() as the expression and Bob’s you uncle (or even Aunt these days) EG produced a unique sequence in the new column.

The Monotonic() function is apparently experimental in SAS 9.1. But was mentioned in a SUGI paper : Helpful Undocumented Features in SAS

And according to Wikipedia it is:

“In mathematics, a monotonic function (or monotone function) is a function which preserves the given order.”

I have to agree with all his points, the only negative I would add is that its integration with the SAS Metadata Server is not where it should be, but it gets better every release, bring on EG 4.2!

EG is a no brainer for any user who has never coded in SAS and a must have for any data analysts worth their salt.

However I have always struggled to persuade seasoned SAS Programmers to make the move. Apart from Graphs, they can always wack out code faster than they can use EG (and I have even tried a few shoot out’s to prove the point).

However with EG projects no being included in Metadata Impact analysis, the improved security capabilities EG allows and the licensing benefits compared to a fully loaded Base SAS PC, corporations are starting to make the decision for them.

Having said that the self proclaimed “SAS Uber Geek” in New Zealand has almost seen the light regarding SAS Data Integration Studio, so maybe he will see some value in EG will be next, and if he does I will be sure to post immediately.