Load Data to LASR :: SAS(R) Data Loader 2.4 for Hadoop: User's Guide

Introduction

Load Data to LASR icon in SAS Data Loader window

Use the Load Data to LASR directive to copy Hadoop tables to a single SAS LASR Analytic Server, or to a grid of SAS LASR Analytic Servers. On the SAS LASR Analytic Servers, you can analyze tables using software such as SAS Visual Analytics.

When you load data onto a single SAS LASR Analytic Server, you configure a connection that is optimized for symmetric multi-processing (SMP). When you load data onto a grid of SAS LASR Analytic Servers, you configure a connection that is optimized for massively parallel processing (MPP).

Note: The Load Data to LASR directive is distinct and separate from the Load to LASR capability that is provided by SAS LASR Analytic Server.

Prerequisites

An administrator must set some special options for SAS Data Loader on the SAS LASR Analytic Server. You must specify a connection in the LASR Analytic Server panel of the Configuration window. For more information about these tasks, see LASR Analytic Servers Panel.

Example

Follow these steps to create and run the Load Data to LASR directive:

In the Directives directives page, click Load Data to LASR.
In the Source Table task, click the schema that contains the source table that you want to load. Clicking the schema displays the tables in that schema. Click the table that you want to load into the SAS LASR Analytic Server software, and then click Next.
In the Target Table task, click the SAS LASR Analytic Server that you want to receive the target table. Clicking displays target table configuration fields and controls.
As needed, change the name in the Target table name field. The field defines the name of the table in the SAS LASR Analytic Server software.
Select options as needed to replace any existing table of the same name or to compress the target table in the SAS LASR Analytic Server software.
Click the Locations link to view or change the default storage options for the target table in the SAS LASR Analytic Server software.
In the Locations window, you can change the SAS folder, the library name, and the required tag that accompanies the table name.
In the Target Table task, click Next.
In the Result task, click Start loading data. SAS proceeds to generate code for the directive and displays the Code icon . Click the icon to open or save the text of the SAS code that comprises the directive.
During the execution of the directive, the Result task displays the Log icon . Click the icon to open or save the SAS log file that is generated during the execution of the directive.
At the conclusion of the directive, the Result banner receives a status icon that indicates the success or failure of the directive. To view the target table on the SAS LASR Analytic Server, click the View Results icon .

Usage Notes

In MapR distributions of Hadoop, massively parallel processing (MPP) is not supported in the LASR procedure. To load data from MapR Hadoop to a SAS LASR Analytic Server, the server definition must assert the SASIOLA option. The SASIOLA option implements symmetric multiprocessing (SMP.) Server definitions are available in the SAS Data Loader Configuration window, in the LASR Analytic Servers panel. For more information about server definitions, see Add or Update Connections to SAS LASR Analytic Servers.

The Load Data to LASR directive moves entire tables. To improve performance, you can filter the rows and manage the columns before you load the table to the SAS LASR Analytic Server. To reduce table size, use the directives Transform Data or Query or Join Data

The Load Data to LASR directive loads Hive tables to a SAS LASR Analytic Server. It does not load HDFS or NFS data directly. This is because the Load Data to LASR directive performs the load using the SAS embedded process. To support the embedded process, the LASR server tag text box in the LASR Server Configuration dialog box is limited to eight characters and must be valid as a SAS libref. These restrictions will not work for HDFS or NFS data because this data requires the server tag to represent the source path in dot-delimited form.