使用R Inline和fortran进行多维数组操作

2024-05-20 • 问答

我想将Fortran代码集成到我的R包中。为了练习，我尝试将inline与Fortran代码一起使用。我可以创建该函数，但是，当我尝试运行它时，出现段错误

library(inline)
src = "
integer i,j,k
do k = 1,prok
  do j = 1,ncolv
    do i = 1,nrowv
        emis(i,k) = veh(i,j) * lkm(i) * ef(j)*pro(k)
     end do
   end do
end do
"
work <- cfunction(sig = signature(
  nrowv="integer",ncolv = "integer",prok = "integer",veh="numeric",lkm = "numeric",ef = "numeric",pro = "numeric",emis = "numeric"),implicit = "none",dim = c("","","(nrowv,ncolv)","(nrowv)","(ncolv)","(prok)",ncolv,prok)"),src,language="F95",verbose = F)

work(nrowv = 3,ncolv = 2,prok = 3,veh = 1:6,lkm = 1:3,ef = 1:2,pro = 1:4,emis = as.numeric(1))

给出：分段故障（核心已转储）我添加了工作功能的图片，以查看子例程的情况。

print(work)
An object of class 'CFunc'
function (nrowv,prok,veh,lkm,ef,pro,emis) 
.Primitive(".Fortran")(<pointer: 0x7ffac77db100>,nrowv = as.integer(nrowv),ncolv = as.integer(ncolv),prok = as.integer(prok),veh = as.double(veh),lkm = as.double(lkm),ef = as.double(ef),pro = as.double(pro),emis = as.double(emis))
<environment: 0x55ab99679e70>
code:
  1: 
  2:  SUBROUTINE file77f62087457e ( nrowv,emis )
  3: IMPLICIT none
  4: INTEGER nrowv
  5: INTEGER ncolv
  6: INTEGER prok
  7: DOUBLE PRECISION veh(nrowv,ncolv)
  8: DOUBLE PRECISION lkm(nrowv)
  9: DOUBLE PRECISION ef(ncolv)
 10: DOUBLE PRECISION pro(prok)
 11: DOUBLE PRECISION emis(nrowv,prok)
 12: 
 13: integer i,k
 14: do k = 1,prok
 15:   do j = 1,ncolv
 16:     do i = 1,nrowv
 17: emis(i,j) * lkm(i) * ef(j)*pro(k)
 18:       end do
 19:     end do
 20: end do
 21: 
 22: RETURN
 23: END
 24:

> work(nrowv = 3L,+ ncolv = 2L,+ prok = 3L,+ veh = as.numeric(1:6),+ lkm = as.numeric(1:3),+ ef = as.numeric(1:2),+ pro = as.numeric(1:3),+ emis = numeric(3*2*3)) $nrowv [1] 3 $ncolv [1] 2 $prok [1] 3 $veh [1] 1 2 3 4 5 6 $lkm [1] 1 2 3 $ef [1] 1 2 $pro [1] 1 2 3 $emis [1] 1 4 9 8 20 36 2 8 18 16 40 72 3 12 27 24 60 108

使用R Inline和fortran进行多维数组操作

ever2003 回答：使用R Inline和fortran进行多维数组操作

大家都在问