如何在聚合R中应用自定义函数

Question

everyone! I have tried to find similar question, but i didn't. Let's assume that we have a simple data frame like this

enter image description here

I need to aggregate by group: aggregate(value~group, data=data, mean,na.rm=TRUE) But, when i use na.rm=TRUE , group 1 disappears. It is absolutely necessary to use na.rm=TRUE , because in opposite case i will get incorrect averages. I also have tried to use custom function:

customMean<-function(x){
  if (all(is.na(x))){
      return (NA_integer_)
  } else {  
    return(mean(x,na.rm=TRUE))
  }
}

but, the result is same. Does anybody know how to fix this issue? I expect that final result will be like this: enter image description here . Thanks

Answer 1

By default, the formula method of aggregate has na.action=na.omit . So, it removes the entire row if there is an NA value. We can change it to na.action=NULL and it should work.

aggregate(value~group, data=data, mean,na.rm=TRUE, na.action=NULL)
#    group value
#1     1   NaN
#2     2   2.5