What’s New in SAS 9.3 In-Database Products

Overview

Starting in SAS 9.3, the user documentation for format publishing, in-database procedures, and the SAS Scoring Accelerator are combined into this document, SAS 9.3 In-Database Products: User’s Guide.

Support for Teradata V13, Netezza V6.0, and Aster nCluster V6 has been added.

Some Base SAS procedures have been enhanced for in-database processing inside Aster nCluster, Greenplum, and Netezza.

In the November 2011 release, format publishing is supported for Aster nCluster and Greenplum. In addition, in-database scoring for Teradata has been enhanced by the addition of the SAS Embedded Process. The SAS Embedded Process is a SAS server process that runs within Teradata to read and write data.

In the December 2011 release, in-database scoring for DB2 has been enhanced by the addition of the SAS Embedded Process.

In the April 2012 release, you can use the SAS Scoring Accelerator in conjunction with SAS Model Manager to manage and deploy scoring models in Greenplum.

In the June 2012 release, in-database scoring is supported for Oracle using the SAS Embedded Process.

In the August 2012 release, in-database scoring for Greenplum has been enhanced by the addition of the SAS Embedded Process. Also, SAS Scoring Accelerator and SAS Model Manager now supports importing SAS/STAT linear models and SAS High-Performance Analytics models from a SAS package file (.SPK).

In the December 2012 release, installation and configuration information for the SAS Embedded Process for Hadoop is added to the SAS In-Database Products: Administrator's Guide. The SAS Embedded Process must be installed and configured before you can read and write data to a Hadoop Distributed File System (HDFS) in parallel for High-Performance Analytics (HPA).

Documentation Enhancements

Starting in SAS 9.3, the user documentation for these in-database technologies has been combined into this document, SAS In-Database Products: User's Guide:

Format publishing and the SAS_PUT( ) function were previously documented in SAS/ACCESS for Relational Databases: Reference.
In-database procedures were previously documented in SAS/ACCESS for Relational Databases: Reference.

Note: Each in-database procedure has its own specific considerations and limitations. For more information, see the documentation for the procedure.
Scoring Accelerator was previously documented in the SAS Scoring Accelerator: User’s Guide for each database.

The in-database installation and configuration documentation can be found in SAS In-Database Products: Administrator's Guide.

The configuration instructions for the SAS Model Manager In-Database Scoring Scripts product have been moved to this book from the SAS Model Manager: User's Guide.

Compiled Publishing Macros

All publishing macros are compiled now for better security. There is no change in how you run the publishing macros.

Additional Alias for INDCONN Macro Password Argument

You can now use PASS= for the password argument in the INDCONN macro variable.

In-Database Procedures

There are several enhancements to in-database procedures:

You can use the SAS In-Database technology to run some Base SAS procedures inside Aster nCluster, Greenplum, and Netezza.
In BY-group processing, the NOTSORTED option is now ignored because the data is always returned in sorted order. Previously, the NOTSORTED option was not supported.

Support for SAS/STAT Linear Models

In the August 2012 release, the SAS Scoring Accelerator now supports importing SAS/STAT linear models and SAS High-Performance Analytics models from a SAS package file (.SPK). Models that have a DATA step score code type can also be scored, published, and included in performance monitoring. SAS Model Manager is required.

Support for Reading and Writing Data to HDFS for High-Performance Analytics

Aster nCluster Changes

The following changes have been made for Aster nCluster:

Support for Aster nCluster V6 has been added.
If you use Aster nCluster V6, you can specify a schema where the scoring model files are published. You specify this schema in the INDCONN macro variable, and you can use the MODEL_SCHEMA parameter in the SAS_SCORE() function when you execute the scoring model.
In the November 2011 release, format publishing is now supported. Format publishing enables you to execute SAS PUT function calls inside the database. You can reference most of the formats that SAS supplies and the custom formats that you create with PROC FORMAT.
In the December 2011 release, the default value for the SQLGENERATION system option now includes Aster nCluster. This means that procedures automatically run inside the database.

DB2 Changes

The following changes have been made for DB2:

Format publishing is now supported. Format publishing enables you to execute SAS PUT function calls inside the database. You can reference most of the formats that SAS supplies and the custom formats that you create with PROC FORMAT.
In the December 2011 release, in-database scoring for DB2 has been enhanced by the addition of the SAS Embedded Process. The SAS Embedded Process is a SAS server process that runs within DB2 to read and write data. The SAS Embedded Process can be used with the SAS Scoring Accelerator for DB2 under UNIX to run scoring models.
In the December 2011 release, the DB2IDA utility was added to control the SAS Embedded Process. DB2IDA is a utility that is installed with the DB2 server. The DB2IDA command enables you to manually stop and restart the SAS Embedded Process without shutting down the database.

Greenplum Changes

The following changes have been made for Greenplum:

In the November 2011 release, format publishing is now supported. Format publishing enables you to execute SAS PUT function calls inside the database. You can reference most of the formats that SAS supplies and the custom formats that you create with PROC FORMAT.
In the December 2011 release, the default value for the SQLGENERATION system option now includes Greenplum. This means that procedures automatically run inside the database.
In the April 2012 release, you can use the SAS Scoring Accelerator in conjunction with SAS Model Manager to manage and deploy scoring models in Greenplum.
In the August 2012 release, in-database scoring for Greenplum has been enhanced by the addition of the SAS Embedded Process. The SAS Embedded Process is a SAS server process that runs within Greenplum to read and write data. The SAS Embedded Process can be used with the SAS Scoring Accelerator for Greenplum to run scoring models.

Note: Starting with the August 2012 release, SAS is not compatible with Greenplum version 4.0 or older when publishing formats or running scoring models. If you use the second maintenance release of SAS 9.3, you must use Greenplum version 4.2.2 or later.

Netezza Changes

The following changes have been made for Netezza:

Support for Netezza V6.0 has been added.
Netezza Performance Server (NPS) is no longer supported.
You can now run Netezza format and model publishing macros in fenced mode and in unfenced mode. Fenced mode means that the format and scoring functions that are published are isolated in a separate process in the Netezza database when they are invoked. An error does not cause the database to stop. When the format or scoring functions are ready for production, you can run the macro to publish the functions in unfenced mode.

Oracle Changes

In the June 2012 release, in-database scoring for Oracle is supported using the SAS Embedded Process. The SAS Embedded Process is a SAS server process that runs within Oracle to read and write data. The SAS Embedded Process can be used with the SAS Scoring Accelerator for Oracle to run scoring models.

Teradata Changes

The following changes have been made for Teradata:

V2R6 on Linux is no longer supported.
In the November 2011 release, in-database scoring for Teradata has been enhanced by the addition of the SAS Embedded Process. The SAS Embedded Process is a SAS server process that runs within Teradata to read and write data. The SAS Embedded Process can be used with the SAS Scoring Accelerator for Teradata to run scoring models. During the installation process, there is an additional RPM file that must be installed. This RPM file contains the SAS Embedded Process. In addition, you must download and install the SAS Embedded Process support functions.