ACCELWHERE= Table Option

Specifies whether to perform WHERE processing optimization by subsetting the data in the Hadoop cluster with MapReduce.

Valid in:	SPD Server
Valid in:	DATA step and PROC step
Requirement:	To perform data subsetting in the Hadoop cluster, there are table and SAS code requirements. For more information, see WHERE Processing Optimization with MapReduce.
Interaction:	The status of the HADOOPACCELWH parameter file option determines the functionality of the ACCELWHERE= table option. See HADOOPACCELWH Parameter File Option .

Syntax

Required Arguments

Details

Example

Syntax

ACCELWHERE=YES | NO

Required Arguments

YES

turns on WHERE processing optimization when the HADOOPACCELWH parameter option is not included (undeclared) in the spdsserv.parm file, which is the default state. If the NOHADOOPACCELWH parameter option is included in the spdsserv.parm file, ACCELWHERE=YES is ignored and WHERE processing optimization is not performed.

NO

turns off WHERE processing optimization when the HADOOPACCELWH parameter option is included (declared) in the spdsserv.parm file.

Details

By default, WHERE processing optimization is not enabled. That is, the spdsserv.parm file does not include the HADOOPACCELWH parameter file option, which requests that data subsetting be performed in the Hadoop cluster with MapReduce. In the default state, you can use the ACCELWHERE=YES table option to turn on WHERE processing optimization in specific cases.

If global WHERE processing optimization is enabled with the HADOOPACCELWH parameter file option, you can use the ACCELWHERE=NO table option to turn off WHERE processing optimization in specific cases.

If WHERE processing optimization is disabled with the NOHADOOPACCELWH parameter file option, the ACCELWHERE= table option setting is ignored.

The ACCELWHERE= table option overrides the SPDSACWH= macro variable. For information about the SPDSACWH= macro variable, see SPDSACWH= Macro Variable.

Example

proc print data=my_lib.my_table (accelwhere=yes)
   where x=1;
run;