Question

df <- data.frame(id = c(1, 2, 3, 3, 3, 4), gender = c("Male", "Female", "Both", "Male", "Female", "Female"))
ids <- unique(df$id)

> df
  id gender
1  1   Male
2  2 Female
3  3   Both
4  3   Male
5  3 Female
6  4 Female

对于每个唯一的id，我想确保如果相应的gender是Both，Male和Female，那么我需要删除与Both对应的行。换句话说，我想要的输出是：

> df
  id gender
1  1   Male
2  2 Female
3  3   Male
4  3 Female
5  4 Female

我试过写一个循环：

按df对id进行子集化，并将每个子集存储到名为sub的列表中
在每个sub内，检查性别是否包含“两个”，“男性”和“女性”
如果是，请删除包含gender =“Both”的行
重新组合data.frame

但是，以下代码并不真正起作用且非常笨重...我想知道在group_by中使用dplyr是否有更简单的方法？

sub <- list()
for(i in 1:length(ids)){
  sub[[i]] <- subset(df, id %in% ids[i])
  if(all(grepl(sub[[i]]$gender, c("Both", "Male", "Female")))){
    sub[[i]] <- sub[[i]][-which(sub[[i]]$gender == "Both"), ]
  }else sub[[i]] = sub[[i]]
}

Answer 1

使用dplyr

df %>% 
    group_by(id) %>% 
    mutate(A = ifelse(length(unique(gender)) >= 3 & gender == 'Both', F, T)) %>% 
    filter(A) %>% 
    select(-A)
# A tibble: 5 x 2
# Groups:   id [4]
     id gender
  <dbl>  <chr>
1     1   Male
2     2 Female
3     3   Male
4     3 Female
5     4 Female

Answer 2

除了tidyverse解决方案，这里还有一个使用lapply的解决方案：

result <- lapply(ids,function(x){
    tmp <- df[df$id == x,]
    if(all(c("Both","Male", "Female") %in% tmp$gender)){
        tmp <- tmp[tmp$gender != "Both",]
    }
    return(tmp)
})
do.call("rbind",result)
#   id gender
# 1  1   Male
# 2  2 Female
# 4  3   Male
# 5  3 Female
# 6  4 Female

R：如何删除循环中子集内的数据

2 个答案: