如何测试NA值是否等于零; 如果是,则替换,如果不是,则保留为NA

Question

编辑为未包含示例数据集中的全部因素,这导致原始解决方案中断.

我正在尝试通过确定是否应该用0替换NA,或者如果将NA替换为NA来清理数据集.

以下是示例数据集.'Dom.Supply'应等于剩余因子的总和.例如,'Feed','Waste','Processing'和'Other.Uses'在行3:5中出现的NA可以用0替换为具有值的因子之和(即'Food'和'种子')等于'Dom.Supply的值.但是,在第1行和第2行中,NAs必须保持为'Food'和'Seed'不等于'Dom.Supply'的总和.

Region  Country Group   Item    Year    Production  Imports Stock.Var   Exports Dom.Supply  Feed    Seed    Waste   Processing  Other.Uses  Food
NAm.Oceania Australia   Cereals Rye 1961    11  0   0   2   9   NA  1   NA  NA  NA  7
NAm.Oceania Australia   Cereals Rye 1962    10  0   0   3   7   NA  1   NA  NA  NA  5
NAm.Oceania Australia   Cereals Rye 1963    10  0   0   1   9   NA  2   NA  NA  NA  7
NAm.Oceania Australia   Cereals Rye 1964    14  0   -5  0   9   NA  2   NA  NA  NA  7
NAm.Oceania Australia   Cereals Rye 1965    11  0   5   0   16  NA  2   NA  NA  NA  14

我曾经replace如下使用这个函数(作为一个例子,但不是我想要使用的操作),但这是用0代替NA的简单替代,而不是NA = 0的测试.

data$AF2 <- 1-((replace(data$Feed, is.na(data$Feed), 0) + (replace(data$Seed,
    is.na(data$Seed), 0)) / data$Dom.Supply))

谢谢!

Answer 1

我用这个:

DF <- read.table(text = "Dom.Supply  Feed    Seed    Waste   Processing  Other.Uses  Food
9   NA  1   NA  NA  NA  7
7   NA  1   NA  NA  NA  5
9   NA  2   NA  NA  NA  7
9   NA  2   NA  NA  NA  7
16  NA  2   NA  NA  NA  14", header = TRUE)

ix <- rowSums(DF[, -1], na.rm = TRUE) == DF[, 1]
DF[ix,] <- lapply(DF[ix,], function(x) {
  x[is.na(x)] <- 0
  x
})
#  Dom.Supply Feed Seed Waste Processing Other.Uses Food
#1          9   NA    1    NA         NA         NA    7
#2          7   NA    1    NA         NA         NA    5
#3          9    0    2     0          0          0    7
#4          9    0    2     0          0          0    7
#5         16    0    2     0          0          0   14