Executing SAS processes in a Hadoop cluster

[juanvg1972]

Hi,

I would like to know which is needed to execute SAS processes in a Hadoop cluster.

I can convert my data setps and proc sql to proc ds2 processes but I want to know if I need some especial licenses in Hadoop to work with SAS.     SAS HPA??, SAS/ACCESS to Hadoop?, SAS in-database? any other? is any of these free ?

Thanks in advance

[AndrewHowell · Accepted solution]

I'm doing the same thing - understanding the relationship between SAS & Hadoop and coming up with best practices to get programmers to use Hadoop effectively without creating horrendous SQL joins with non-Hadoop tables & causing massive I/O pipeline blockages..

I suggest you review the following documentation:

• SAS and Hadoop technology: Deployment Scenarios (to understand your infrastructure)
• SAS 9.4 In-Database Products: Administrator's Guide (especially Ch 5: Administrator's Guide for Hadoop)
• SAS and Hadoop Technology: Overview (especially Ch 4: What SAS Technology interacts with Hadoop?)
• SAS LASR Analytic Server 2.7: Reference Guide (which explains the in-memory procedures)

None of the software is "free" - there is some Hadoop functionality which comes with Base/SAS, but ideally you'd require at least the SAS/Access to Hadoop engine, and - depending on your requirements - possibly some in-database accelerators.

[ChrisNZ]

My (limited) understanding is that in-database processing requires its namesake license, which is apparently now par of SAS® Data Loader for Hadoop.

SAS/ACCESS to Hadoop is the interface to Hive. 

SAS/ACCESS to Impala is useful too to speed up queries.  

None of these are free.