如何在 R 中进行 FOR 和 IF 循环以根据两个因子级别对值求和并创建新变量？

Question

我希望我的数据框像

> trial_acc
   df$Condition df$id df$correct keep
1            1_1 006          1   no
2            3_3 006          1   no
3            1_3 006          5   no
4            3_1 006          3   no
5            1_1 0ui         13   yes
6            3_3 0ui         18   yes
7            1_3 0ui         16   yes
8            3_1 0ui         13   yes
9            1_1 12f         12   no
10           3_3 12f          2   no
11           1_3 12f         13   no
12           3_1 12f          7   no

如果 df$correct 的总数 >= 40 按 df$id 分组，我需要创建一个变量“keep”，其中“是”，否则为“否”

我试过这个，但它不起作用

for (i in levels(trial_acc$`df$id`)) {
  if (trial_acc$`df$id` == i & sum(trial_acc$`df$correct`) >=  40){
    trial_acc$keep <- "yes"
  } else{
    trial_acc$keep <- "no"
  }
}

Answer 1

通过trail_acc$keep <- "yes"或"no" ，您正在更改整个列的值。 要仅更改受影响行的值，我将使用以下方式：

trail_acc$keep <- 'no'  # set default to "no"
for(i in unique(trail_acc$id)){
    if(sum(trail_acc$correct[trail_acc$id == i]) >= 40){
        trail_acc$keep[trail_acc$id == i] <- 'yes'
    }
}

您还可以使用ave和ifelse以矢量化方式进行ifelse ：

sum_by_id <- ave(trail_acc$correct, trail_acc$id, FUN = sum)
trail_acc$keep <- ifelse(sum_by_id >= 40, 'yes', 'no')

#    Condition  id correct keep
# 1        1_1 006       1   no
# 2        3_3 006       1   no
# 3        1_3 006       5   no
# 4        3_1 006       3   no
# 5        1_1 0ui      13  yes
# 6        3_3 0ui      18  yes
# 7        1_3 0ui      16  yes
# 8        3_1 0ui      13  yes
# 9        1_1 12f      12   no
# 10       3_3 12f       2   no
# 11       1_3 12f      13   no
# 12       3_1 12f       7   no

注意我使用了带有简化列名的trail_acc ：

trail_acc <- structure(list(Condition = structure(c(1L, 4L, 2L, 3L, 1L, 4L, 
2L, 3L, 1L, 4L, 2L, 3L), .Label = c("1_1", "1_3", "3_1", "3_3"
), class = "factor"), id = structure(c(1L, 1L, 1L, 1L, 2L, 2L, 
2L, 2L, 3L, 3L, 3L, 3L), .Label = c("006", "0ui", "12f"), class = "factor"), 
    correct = c(1L, 1L, 5L, 3L, 13L, 18L, 16L, 13L, 12L, 2L, 
    13L, 7L)), row.names = c("1", "2", "3", "4", "5", "6", "7", 
"8", "9", "10", "11", "12"), class = "data.frame")

Answer 2

如果你更喜欢dplyr ，你可以这样做：（我使用@mt1022 的数据和简化列）

trail_acc%>%group_by(id)%>%
  mutate(keep = ifelse(sum(correct) >= 40, 'yes', 'no'))%>%
  ungroup()

    # A tibble: 12 x 4
   Condition id    correct keep 
   <fct>     <fct>   <int> <chr>
 1 1_1       006         1 no   
 2 3_3       006         1 no   
 3 1_3       006         5 no   
 4 3_1       006         3 no   
 5 1_1       0ui        13 yes  
 6 3_3       0ui        18 yes  
 7 1_3       0ui        16 yes  
 8 3_1       0ui        13 yes  
 9 1_1       12f        12 no   
10 3_3       12f         2 no   
11 1_3       12f        13 no   
12 3_1       12f         7 no