The SORTDUP= option
specifies what variables to sort to remove duplicate observations
when the SORT procedure NODUPRECS option is specified.
When SORTDUP= is set
to LOGICAL and NODUPRECS is specified in the SORT procedure, duplicate
observations are removed based on the variables that remain after
a DROP or KEEP operation on the input data set. Setting SORTDUP=LOGICAL
increases the number of duplicate observations that are removed because
it eliminates variables before observations are compared. Setting
SORTDUP=LOGICAL might improve performance.
When SORTDUP= is set
to PHYSICAL and NODUPRECS is specified in the SORT procedure, duplicate
observations are removed based on all of the variables in the input
data set.