Automate using a function in R-CodePudding

I have a dataset where I need to do the following steps for analysis.

a <- with(iris, sum(Petal.Length[Petal.Width<0.2]))
b <- with(iris, sum(Petal.Length[Petal.Width<0.7]))
cc <- with(iris, sum(Petal.Length[Petal.Width<1]))
d <- with(iris, sum(Petal.Length[Petal.Width<3]))

e<-(b-a)
f<-(cc-b)
g<-(d-cc)

newdata <- data.frame(x=c(1,2,3),y=c(e,f,g))

I now need to calculate the same by replacing Petal.Length to Sepal.Length and so on. Therefore I wrote a function for this but I get the error 'Sepal.Length' not found when using h<- myfun(iris,Sepal.Length). Any help is appreciated.

myfun <- function(mydata,myvar){
  res <- NULL
  a <- with(mydata, sum(myvar[Petal.Width<0.2]))
  b <- with(mydata, sum(myvar[Petal.Width<0.7]))
  cc <- with(mydata, sum(myvar[Petal.Width<1]))
  d <- with(mydata, sum(myvar[Petal.Width<3]))
  
  e<-(b-a)
  f<-(cc-b)
  g<-(d-cc)
  
  res <- c(e,f,g)
  
}

CodePudding user response：

This is a simple fix. Remember to put the return in the function. When you pass input arguments, myvar should be in the form of character

myfun <- function(mydata,myvar){
  a = sum(mydata[mydata[,"Petal.Width"] < 0.2 ,myvar])
  b = sum(mydata[mydata[,"Petal.Width"] < 0.7 ,myvar])
  c = sum(mydata[mydata[,"Petal.Width"] < 1 ,myvar])
  d = sum(mydata[mydata[,"Petal.Width"] < 3 ,myvar])
  
  e<-(b-a)
  f<-(c-b)
  g<-(d-c)
  
  return(c(e,f,g))
  
}

myfun(iris,"Petal.Length")

If you want to write it using the functions already present in R you can use cut which divides a numerical variable into intervals, and then via sapply adds Petal.Length for each interval.

lev = c(0.2,0.7,1,3)
cutPoint = cut(iris$Petal.Width,
               breaks = c(min(iris$Petal.Width),0.2,0.7,1,3),
               labels = lev,
               right = F)
sapply(1:4, function(i) sum(iris$Petal.Length[cutPoint == lev[i]]))

CodePudding user response：

Try

myfun <- function(mydata,myvar){
  a <- sum(mydata[myvar]<0.2)
  b <- sum(mydata[myvar]<0.7)
  cc <- sum(mydata[myvar]<1)
  d <- sum(mydata[myvar]<3)
  
  e<-(b-a)
  f<-(cc-b)
  g<-(d-cc)
  
  c(e,f,g)
}

myfun(iris,"Petal.Length")

results in

[1]  0  0 50