在R中编写函数，收到错误“eval中的错误（expr，envir，enclos）：...”

Question

writing code for homework. The code goes in fine, but when I try the function, I receive this #'ed error (which I do not understand). The code is supposed to take a dataframe of coordinates, keep the coords, and add columns denoting which hemisphere the coordinates is part of. So the question is how do I change the function/code to get rid of the error?

Also, the code uses packages sp and foreach.

Here's the dataframe input and my errored output:

set.seed(10)
n <- 10
df <- data.frame(xpos=runif(n,0,360),ypos=runif(n,-90,90))
df

outHemisphere <- hemisphereSummary(df=df)
outHemisphere
# Error in eval(expr, envir, enclos) : argument is missing, with no default

Here's what the output is supposed to be (after the set seed and df are in workspace):

outHemisphere <- hemisphereSummary(df=df)
outHemisphere
#           coordinates EWhemisphere NShemisphere
#1    (182.692, 27.298)            W            N
#2   (110.437, 12.1928)            E            N
#3  (153.687, -69.5684)            E            S
#4   (249.517, 17.2666)            W            N
#5   (30.6489, -25.551)            E            S
#6  (81.1572, -12.8143)            E            S
#7   (98.831, -80.6574)            E            S
#8   (98.0298, -42.448)            E            S
#9  (221.699, -18.2177)            W            S
#10  (154.682, 60.5041)            E            N

Here's my code/functions for the assignment, which inputs without errors:

    hemisphereSummary <- function(df, projargs="+proj=longlat +datum=WGS84")
  {

  # install.packages("foreach")
  # install.packages("sp")
  library("foreach")
  library("sp")
  if(class(df)!="data.frame") stop ("df must be a data frame.") 
  registerDoSEQ() # register the non-parallel backend for foreach.

  df_mat <- cbind(df[,1], df[,2]) #makes matrix of 2 cols, drawing from df.
  row.names(df_mat) <- as.character(1:nrow(df_mat))
  df_CRS <- CRS(projargs) # correct projection arguments.
  df_sp <- SpatialPoints(coords=df_mat,proj4string=df_CRS) # makes spatial object!

  #  df_sp


  foreachloop <- foreach(i = 1:(nrow(df)), .packages = "sp", .combine="rbind",) %dopar% 
 { 
   findhemisphere <- function(i, df_sp, df) # nested function to sort which hemi.
   {
     coords <- coordinates(df_sp[i,])
     hemi_names <- df    
     colnames(hemi_names) <- c("EWhemisphere", "NShemisphere")
     #if-else's: E=0-180, W=181-360. N=0 to +90, S=-1 to -90.
     ifelse(coords[1] >=0 & coords[1] <=180,coords[1] <- "E", coords[1] <- "W")
     ifelse(coords[2] >=0 & coords[2] <=90,coords[2] <- "N", coords[2] <- "S")
     return(hemi_names)[i,] <- coords    
   } 
 }

  SPdataframe <- SpatialPointsDataFrame(coords=coordinates(df_sp), 
  data = foreachloop, proj4string = df_CRS,  match.ID = FALSE)
  }

Answer 1

The error:

Error in eval(expr, envir, enclos) : argument is missing, with no default

is caused by the trailing comma in the call to foreach since that is interpreted as a missing argument to foreach . But you'll run into more problems after you remove it. For example, the foreach loop is defining, not calling, the same function repeatedly and combining those functions with rbind . Presumably you want to define the function outside the loop, call it in the loop, and combine the results with rbind .

I'm also skeptical about the last line of the foreach loop:

return(hemi_names)[i,] <- coords

But I don't think there's any reason to use foreach for this. I believe the coordinates function can operate on the entire object df_sp , so there's no reason for a loop at all.