Hi,
I've implemented lazy versions of them in Rcpp::sugar.
Here are some examples:
fx <- cxxfunction( signature( x = "integer" ), '
IntegerVector xx(x) ;
List res = lapply( xx, seq_len );
return res ;
', plugin = "Rcpp" )
checkEquals( fx( 1:10 ), lapply( 1:10, seq_len ) )
fx <- cxxfunction( signature( x = "numeric" ), '
NumericVector xx(x) ;
return List::create(
pmin( xx, 5),
pmin( 5, xx)
) ;
', plugin = "Rcpp" )
checkEquals( fx(1:10) ,
list(
c(1:5,rep(5,5)),
c(1:5,rep(5,5))
)
)
fx <- cxxfunction( signature( x = "numeric", y = "numeric" ), '
NumericVector xx(x) ;
NumericVector yy(y) ;
NumericVector res = ifelse( xx < yy, xx*xx, -(yy*yy) ) ;
return res ;
', plugin = "Rcpp" )
x <- 1:10
y <- 10:1
checkEquals( fx( x, y), ifelse( x<y, x*x, -(y*y) ) )
One thing to notice about ifelse is that it only evaluates either the
rhs or the lhs, given the condition. This differs from R which has to
evaluate both.
Romain