What’s New In SAS LASR Analytic Server

Overview

SAS LASR Analytic Server 2.4 includes the following changes:
  • Enhancements to the IMSTAT procedure
  • Compression for in-memory tables and SASHDAT tables
  • Enhancements to the RECOMMEND procedure
  • Documentation enhancements

Enhancements to the IMSTAT Procedure

The following enhancements have been made to the IMSTAT procedure:
  • The AGGREGATE statement is new for the 2.4 release. You can aggregate values of one or more variables and specify the aggregate method to use for each variable.
  • The CLUSTER statement is enhanced with a CODE option. You can use this option to save the DATA step scoring code for the clustering.
  • The CREATETABLE statement is new. It is used to create an empty in-memory table that is later used for appending records to it. It had been added to support working with event stream processing.
  • The FORECAST statement is enhanced to support goal-seeking forecasts. The CONTROLVARS=, UPPER=, and LOWER= options are used to specify the control variables and bounds. The GOAL= option is used to specify the goal variable and the WEIGHT= option is used to specify a weight for the goal variable.
  • The GENMODEL statement is enhanced with the VARSELECTION option to support variable selection.
  • The GLM statement is enhanced with the VIF option to support computing variable inflation factors and tolerances.
  • The HISTOGRAM statement is enhanced with the ROUNDINGDIRECTION= option.
  • The SLSTAY= option in the GLM, LOGISTIC, and GENMODEL statements automatically triggers variable selection. In previous releases, you had to specify the VARSELECTION or SHOWSELECTION option to trigger variable selection.

Compressed In-Memory Tables and SASHDAT Tables

The server can read and write compressed in-memory tables. You can compress data as you load data to memory as follows:
  • The SAS LASR Analytic Server engine is enhanced with a SQUEEZE= option. The server compresses the data as the table is loaded to memory.
  • The LASR procedure is enhanced to support a SQUEEZE option. When you specify the option, the server compresses the data as the table is loaded to memory.
When a table is already in-memory, you can compress it with the COMPRESS statement in the IMSTAT procedure.
SASHDAT tables in HDFS can also be compressed. The advantage of compressing tables in HDFS is not to save disk space, but to combine the in-memory table compression with the memory efficiencies of the SASHDAT format. You can use SASHDAT tables with compression as follows:
  • The SAS Data in HDFS engine is enhanced with a SQUEEZE= option. The data are compressed as the rows are added to the SASHDAT table in HDFS.
  • The SAVE statement for the IMSTAT procedure is enhanced with a SQUEEZE option that saves the in-memory table to HDFS in compressed form.

Enhancements to the RECOMMEND Procedure

The PREDICT statement is enhanced to include a TEMPTABLE option and the ITEMDATA= option.

Enhancements to Data Access

For distributed SAS LASR Analytic Server, the syntax for loading data in parallel from remote data clusters is simplified:
  • You no longer need to specify the GRIDDATASERVER= option in the PERFORMANCE statement or set the GRIDDATASERVER environment variable.
  • You no longer need to specify the GRIDMODE= option in the PERFORMANCE statement or set the GRIDMODE environment variable.
This enhancement is common to SAS high-performance procedures. For more information, see "Data Access Modes" in Base SAS Procedures Guide: High-Performance Procedures.

Documentation Enhancements

The following changes are included in this release:
  • The ORDERBYDESC option for the SUMMARY statement of the IMSTAT procedure was incorrectly identified as the ORDERDESC option in previous releases of this document.
  • Two examples of the FORECAST statement in the IMSTAT procedure are added.
  • Information about working with HDFS has been added. See Common HDFS Commands.
  • Information about passwordless SSH has been enhanced and moved to a single section. See Passwordless SSH.
  • Information about the IMPORTCUBE statement in the IMSTAT procedure is removed.
Copyright © SAS Institute Inc. All rights reserved.