In order to run a DATA
step program in Hadoop, the following is required:
-
The DSACCEL= system option is set
to ANY.
For more information
about the DSACCEL= system option, see SAS System Options: Reference.
-
The code must contain a LIBNAME
statement using the SAS/ACCESS HADOOP engine.
For more information
about the Hadoop LIBNAME statement, see SAS/ACCESS for
Relational Databases: Reference.
-
The input and output files must
use the same libref for the HADOOP engine.
-
The DATA statement must be followed
immediately by the SET statement.
This example demonstrates
these requirements:
options dsaccel=any;
libname hdone hadoop;
data hdone.out;
set hdone.in;
/* DATA step code */
run;
-
The SAS Embedded Process must be
running on the cluster where the input and output files exist.