r - 如何在函数调用中使用 data.table？

Question

1

我想用data.table在自定义函数中，我收到错误:

> library(data.table)
> DT = data.table(a = c(6,3,1,9,NA), b = 4:8)
> DT
    a b
1:  6 4
2:  3 5
3:  1 6
4:  9 7
5: NA 8
> fillna = function(df,var){
+   col = df[[var]]
+   set(df, i = which(is.na(col)), j = var, value = mean(col, na.rm=T))
+ }
>
> fillna(DT,a)
Error in (function(x, i, exact) if (is.matrix(i)) as.matrix(x)[[i]] else .subset2(x,  :
  object 'a' not found
>

如何解决？

2

我也尝试使用 get()从这篇文章:

Making function with data.table in R

但它仍然失败:

> fillna = function(df,var){
+   df[is.na(get(var)),var:=mean(df[[var]],na.rm = T)]
+   df
+ }
>
> fillna(DT,a)
Error in .checkTypos(e, names_x) :
  Object 'a' not found. Perhaps you intended a, b

Answer 1

一个简单的解决方法是将列名作为字符串传递

fillna = function(df,var){
    col = df[[var]]
    set(df, i = which(is.na(col)), j = var, value = mean(col, na.rm=T))
    return(df)
}

fillna(DT,"a")

#      a b
#1: 6.00 4
#2: 3.00 5
#3: 1.00 6
#4: 9.00 7
#5: 4.75 8