R使用字符串来检索数据列

Question

I am struggling with something that looks simple; but I am stuck on it for quite some time now.

I have a data.frame quite long but here's a sample that would represent it.

my.dataframe<-data.frame(PointA.X=sample(100,4))
my.dataframe$PointA.Y<-sample(100,4)
my.dataframe$PointB.X<-sample(100,4)
my.dataframe$PointB.Y<-sample(100,4)

     PointA.X PointA.Y PointB.X PointB.Y
1       93       98       46       45
2       58        3       80       89
3       61       64       17       14
4       56       46       65       23

I am looking for making a function that would use two arguments from which more would be arguments would arise.

MyFunction<-function(Start, End){
XStart <- get(as.character(paste0("Mydataframe$" , Start , ".X")))
XEnd   <- get(as.character(paste0("Mydataframe$" , End   , ".X")))
YStart <- get(as.character(paste0("Mydataframe$" , Start , ".Y")))
YEnd   <- get(as.character(paste0("Mydataframe$" , End   , ".Y" )))
sqrt(((XStart - XEnd) ^ 2 + (YStart - YEnd) ^ 2))
} # End of My Function

In this case I would define the StartPoint and the EndPoint to calculate the length of a segment between them. MyFunction("PointA", "PointB")

To my understanding in

MyFunction("PointA", "PointB")

the following

as.character(paste0("Mydataframe$" , Start , ".X")) 

returns

"Mydataframe$PointA.X"

which is a valid column in my dataframe Using get() is looking for an object instead of looking for the actual data.

That's where I am stuck. Is there a function for returning to the value I am looking for?

Thank you all in advance

Answer 1

Try this. It may help.


MyFunction<-function(Start, End){
XStart <- eval(parse(text=paste("my.dataframe$",Start,".X", sep = "")))
XEnd   <- eval(parse(text=paste("my.dataframe$",End,".X", sep = "")))
YStart <- eval(parse(text=paste("my.dataframe$",Start,".Y", sep = "")))
YEnd   <- eval(parse(text=paste("my.dataframe$",End,".Y", sep = "")))
sqrt(((XStart - XEnd) ^ 2 + (YStart - YEnd) ^ 2))
}

Answer 2

As suggested by Richard, it is possible to subset a data frame using a string into the brackets [[]] but not with the $ symbol.

So advice for the future: use brackets...

  MyFunction<-function(Start, End){
  XStart <- my.dataframe[[paste0(Start, ".X")]]
  YStart <- my.dataframe[[paste0(Start, ".Y")]]

  XEnd <- my.dataframe[[paste0(End, ".X")]]
  YEnd <- my.dataframe[[paste0(End, ".Y")]]

  sqrt(((XStart - XEnd) ^ 2 + (YStart - YEnd) ^ 2))
} # End of My Function

MyFunction("PointA", "PointB") # Note the arguments are provided as characters
> [1] 39.20459 80.52950 34.17601  6.00000

More interesting I can also loop the function across the column names. So if the datagrams is longer.

my.dataframe<-data.frame(PointA.X=sample(100,4))
my.dataframe$PointA.Y<-sample(100,4)
my.dataframe$PointB.X<-sample(100,4)
my.dataframe$PointB.Y<-sample(100,4)
my.dataframe$PointC.X<-sample(100,4)
my.dataframe$PointC.Y<-sample(100,4)

And the function remains the same:

MyFunction<-function(Start, End){
XStart <- my.dataframe[[paste0(Start, ".X")]]
YStart <- my.dataframe[[paste0(Start, ".Y")]]

XEnd <- my.dataframe[[paste0(End, ".X")]]
YEnd <- my.dataframe[[paste0(End, ".Y")]]

sqrt(((XStart - XEnd) ^ 2 + (YStart - YEnd) ^ 2))
} # End of My Function

I can build a for loop :

for (VariableI in seq(from=1, to=length(colnames(my.dataframe)), by=2)){
Start<-unlist(strsplit(colnames(my.dataframe)[VariableI], "[.]"))[1]
End<-unlist(strsplit(colnames(my.dataframe)[VariableI+2], "[.]"))[1]
assign(paste0(Start,End), MyFunction(Start, End)) 
}

Create the following objects

 PointAPointB
    [1] 32.57299 74.30343 73.08215 83.25863
    PointBPointC
    [1]  5.385165 90.609050 68.883960 58.137767

I guess I am just missing PointAPointC. I might use the combine function to walk around this

 combn(colnames(my.dataframe), 2)

    [,1]       [,2]       [,3]       [,4]       [,5]       [,6]       [,7]       [,8]       [,9]       [,10]     
[1,] "PointA.X" "PointA.X" "PointA.X" "PointA.X" "PointA.X" "PointA.Y" "PointA.Y" "PointA.Y" "PointA.Y" "PointB.X"
[2,] "PointA.Y" "PointB.X" "PointB.Y" "PointC.X" "PointC.Y" "PointB.X" "PointB.Y" "PointC.X" "PointC.Y" "PointB.Y"
     [,11]      [,12]      [,13]      [,14]      [,15]     
[1,] "PointB.X" "PointB.X" "PointB.Y" "PointB.Y" "PointC.X"
[2,] "PointC.X" "PointC.Y" "PointC.X" "PointC.Y" "PointC.Y"