R 按条件找到 NA 个单元格，然后按其他条件赋值

Question

我有一个像这样的

df

，我想通过非常复杂的条件来填充

NA

，但我不知道编码..

df <- data.frame(
    age  = c("age15","age15","age16","age16"),
    occo = c("ag","ag","uuse","ag"),
    occn = c("tw","uuse","use","uuse"),
    num  = c(12,NA,567,NA),
    occo2 = c("ag","use","ag","use"),
    occn2 = c("tw","use","tw","use"), 
    num2  = c(2,45,67,789)
)

首先，我想通过条件

NA

找到

num

列中的

occo == "ag" & occn == "uuse"

单元格，然后以条件

num2

为

occo2 == "ag" & occn2 == "tw"

赋值，即

2

。 更重要的是，按

age

组，因此第二个

NA

将被

67

取代。

预期结果将是：

df <- data.frame(
    age  = c("age15","age15","age16","age16"),
    occo = c("ag","ag","uuse","ag"),
    occn = c("tw","uuse","use","uuse"),
    num  = c(12,2,567,67),
    occo2 = c("ag","use","ag","use"),
    occn2 = c("tw","use","tw","use"), 
    num2  = c(2,45,67,789)
)

类似这样的事情

df2 <- df %>%
 group_by(age) %>%
 mutate(num = case_when(
  occo == "ag" & occn == "uuse" ~ .[which(occo2 == "ag", occn2 == "tw"),][7]
))

Answer 1

这会创建所需的输出，但显然不是预期的方式：

df |> 
mutate(num = if_else(is.na(num), lag(num2), num))

我对规则的假设是：

对于每个年龄段，有一个 num2 为 occo2 == "ag" & occn2 == "tw"，还有一个 num 为 NA，其中 occo == "ag" & occn == "uuse"。将每个组的 num 值设置为来自同一组且满足上述条件的 num2 值。不过还不清楚