We’re smarter together. Learn from this collection of community knowledge and add your expertise.

How to back up SAS Enterprise Miner projects without data sets

by Community Manager Thursday (579 Views)

Question

How do you back up SAS Enterprise Miner projects/diagrams without data sets? Currently I manually save diagrams as xml files or create .spk files.

Answer

Contributed by David Ogden, SAS Sr. Manager, Analytical Consulting, and Dominique Latour, SAS Distinguished Research Statistician Developer

 

For flows that are “large”…or have lots of metadata…I’ve experienced an overflow issue when exporting as xml. 

There is a more robust way to back-up a flow, while also not backing-up any of the data created by the EM project:  saving your flow as a batch program.

Here are the steps:

  1. From inside SAS Enterprise Miner, ensure that all nodes terminate in a single node.  You can use a Control Point node to accomplish this.
  2. Right-click on the terminal node (e.g. the Control Point node), and select to “Export Path as SAS Program.”
  3. Deselect both options that are given; navigate to the location where you want to save the batch SAS Program, and name the program.  For example, flow123_backup.sas.


This large SAS program, if executed, would recreate all the folders, subfolders, and metadata necessary to restore the diagram.


Before executing this program, to restore the project, there are a couple of places that you need to modify this batch program:

%let EM_PROJECT =;   
<-- set this to the path where you want the project to get restored to.  (by default it goes in WORK)

%let EM_PROJECTNAME =; 
<-- set this to the name of the project (by default it’s EmProject)



So, if you wanted the restored project to be in a folder called “flow123” and to be located in the UNIX directory of /gpfs/sasdata/, you would have to modify 2 lines of code found in the flow123_backup.sas batch program as:

%let EM_PROJECT=/gpfs/sasdata;

%let PROJECTNAME=flow123;


Not only is none of the data created by the EM flow restored…but neither is the input dataset.  So, after you open this restored project, you’d need to point to the modeling data source, drop it in the project, and re-run the flow.

 

Contributors
Your turn
Sign In!

Want to write an article? Sign in with your profile.