CODEWITHSUNDEEP
rfunctionobjectloops
Joschi
Joschi

Reputation: 3039

Loop within a function and automatically create objects in R

I try to calculate the column means for diffrent groups in R. there exist several methods to assign groups and so two columns where created that contain diffrent groupings.

    # create a test df
    df.abcd.2<-data.frame(Grouping1=c("a","f","a","d","d","f","a"),Grouping2=c("y","y","z","z","x","x","q"),Var1=sample(1:7),Var2=sample(1:7),Var3=rnorm(1:7))
    df.abcd.2

Now I created a loop with assign, lapply, split and colMeans to get my results and store the in diffrent dfs. The loop works fine.

   #Loop to create the colmeans and store them in dataframes
    for (i in 1:2){
      nam <- paste("RRRRRR",deparse(i), sep=".")
      assign(nam, as.data.frame(
        lapply(
          split(df.abcd.2[,3:5], df.abcd.2[,i]), colMeans)
      )
      )
    }

So now i would like to create a function to apply this method on diffrent dataframes. My attemp looked like this:

    # 1. function to calculate colMeans for diffrent groups
    # df= desired datatframe, 
    # a=starting column: beginning of the columns that contain the groups, b= end of columns that contain the groups
    # c=startinc column: beginning of columns to be analized, d=end of columns do be analized

    function.split.colMeans<-function(df,a,b,c,d)
    {for (i in a:b){
      nam <- paste("OOOOO",deparse(i), sep=".")
      assign(nam, as.data.frame(
        lapply(
          split(df[,c:d], df[,i]), colMeans)
      )
      )
    }
    }
    #test the function
    function.split.colMeans(df.abcd.2,1,2,3,5)

So when I test this function I get neither an error message nor results... Can anyone help me out, please?

Upvotes: 1

Views: 4435

Answers (1)

Spacedman
Spacedman

Reputation: 94202

It's working perfectly. Read the help for assign. Learn about frames and environments.

In other words, its creating the variables inside your function, but they don't leak out into the environment you see when you do ls() at the command line. If you put print(ls()) inside your functions loop you'll see them, but when the function ends, they disappear.

Normally, the only way functions interact with their calling environment is by their return value. Any other method is entering a whole world of pain.

DONT use assign to create things with sequential or informative names. Ever. Unless you know what you are doing, which you don't... Stick them in lists, then you can index the parts for looping and so on.

Upvotes: 1

Related Questions