我要根据类别2在类别3和类别4中是否具有Y来创建类别6的以下数据。即如果Category3 == Y，则Category6 == Cat3；如果Category4 == Y，则Category6 == Cat4。
由于Orange和Grape的Category3和Category4都为Y，所以我想复制该行，使该行中的所有其他内容保持不变：
拥有：
Category1 Category2 Category3 Category4 Category5
Phase 1 Apple Y N 5
Phase 1 Berry N Y 3
Phase 2 Orange Y Y 4
Phase 2 Grape Y Y 2
想要
Category1 Category2 Category3 Category4 Category5 Category6
Phase 1 Apple Y N 5 Cat3
Phase 1 Berry N Y 3 Cat4
Phase 2 Orange Y Y 4 Cat3
Phase 2 Orange Y Y 4 Cat4
Phase 2 Grape Y Y 2 Cat3
Phase 2 Grape Y Y 2 Cat4
这里是一个示例，我发现它与我的非常相似，但是由于要尝试保留其他变量，因此我似乎无法使其正确地适合我的数据。我有什么办法可以保留这些？
Duplicating Rows by creating unique columns
复制：
dat <- data.frame(Category1 = c("Phase 1", "Phase 1", "Phase 2", "Phase 2"),
Category2 = c("Apple", "Berry", "Orange", "Grape"),
Category3 = c("Y", "N", "Y", "Y"),
Category4 = c("N", "Y", "Y", "Y"),
Category5 = c("5", "3", "4", "2"))
答案 0 :(得分：3)
这是基本的R方法。 dat来自Tyler's answer。
nm = c("Category3", "Category4")
ind = which(dat[nm] == "Y", arr.ind = TRUE)
ind = ind[order(ind[,1], ind[,2]),]
transform(dat[ind[,1],], Category6 = nm[ind[,2]])
# Category1 Category2 Category3 Category4 Category5 Category6
#1 Phase 1 Apple Y N 5 Category3
#2 Phase 1 Berry N Y 3 Category4
#3 Phase 2 Orange Y Y 4 Category3
#3.1 Phase 2 Orange Y Y 4 Category4
#4 Phase 2 Grape Y Y 2 Category3
#4.1 Phase 2 Grape Y Y 2 Category4
答案 1 :(得分：2)
使用tidyverse，您可以使用case_when()动词来完成此操作，如下所示：
library(tidyverse)
dat <- data.frame(
Category1 = c("Phase 1", "Phase 1", "Phase 2", "Phase 2"),
Category2 = c("Apple", "Berry", "Orange", "Grape"),
Category3 = c("Y", "N", "Y", "Y"),
Category4 = c("N", "Y", "Y", "Y"),
Category5 = c("5", "3", "4", "2")
)
# First create a variable that identifies which of Category3 or Category4 was "Y"
dat %>%
mutate(Category6 = case_when(
Category3 == "Y" ~ "Cat3",
Category4 == "Y" ~ "Cat4"
)) -> dat
# Next duplicate rows where Category3 and Category4 are both "Y"
# but since the previous option encoded Category 6 to "Cat3" first,
# here we will give the new duplicate a Category6 value of "Cat4"
dat2 <- dat
for(i in 1:nrow(dat)){
new_dat <- dat[i,]
if(new_dat$Category3 == "Y" & new_dat$Category4 == "Y"){
new_dat$Category6 <- "Cat4"
dat2 <- rbind(dat2, new_dat)
}
}
输出：
Category1 Category2 Category3 Category4 Category5 Category6
Phase 1 Apple Y N 5 Cat3
Phase 1 Berry N Y 3 Cat4
Phase 2 Orange Y Y 4 Cat3
Phase 2 Grape Y Y 2 Cat3
Phase 2 Orange Y Y 4 Cat4
Phase 2 Grape Y Y 2 Cat4
答案 2 :(得分：1)
这是将apply与dplyr和tidyr结合使用的一种方法-
dat %>%
mutate(
Category6 = apply(.[3:4], 1 , function(x) names(.[3:4])[x == "Y"])
) %>%
unnest()
Category1 Category2 Category3 Category4 Category5 Category6
1 Phase 1 Apple Y N 5 Category3
2 Phase 1 Berry N Y 3 Category4
3 Phase 2 Orange Y Y 4 Category3
4 Phase 2 Orange Y Y 4 Category4
5 Phase 2 Grape Y Y 2 Category3
6 Phase 2 Grape Y Y 2 Category4
答案 3 :(得分：1)
这里是使用tidyverse
library(dplyr)
library(tidyr)
library(stringr)
dat %>%
gather(key, val, Category3:Category4) %>%
group_by(Category2) %>%
mutate(Category6 = toString(str_replace(key, "(^.{3})\\D+",
"\\1")[val == "Y"])) %>%
spread(key, val) %>%
separate_rows(Category6) %>%
select(names(dat), Category6)
# A tibble: 6 x 6
# Groups: Category2 [4]
# Category1 Category2 Category3 Category4 Category5 Category6
#* <fct> <fct> <chr> <chr> <fct> <chr>
#1 Phase 1 Apple Y N 5 Cat3
#2 Phase 1 Berry N Y 3 Cat4
#3 Phase 2 Grape Y Y 2 Cat3
#4 Phase 2 Grape Y Y 2 Cat4
#5 Phase 2 Orange Y Y 4 Cat3
#6 Phase 2 Orange Y Y 4 Cat4
或者另一个选择是pmap和unnest
library(purrr)
dat %>%
mutate(Category6 = select(., Category3:Category4) %>%
pmap(~ names(which(c(...) == "Y")) %>%
str_replace("^(...)\\D+", "\\1") )) %>%
unnest(Category6)
dat <- structure(list(Category1 = c("Phase 1", "Phase 1", "Phase 2",
"Phase 2"), Category2 = c("Apple", "Berry", "Orange", "Grape"
), Category3 = c("Y", "N", "Y", "Y"), Category4 = c("N", "Y",
"Y", "Y"), Category5 = c("5", "3", "4", "2")),
class = "data.frame", row.names = c(NA,
-4L))
答案 4 :(得分：1)
这使用gather和inner_join来提供笛卡尔积。
library(tidyr)
library(dplyr)
half_result <- gather(dat,key, value, c('Category3', 'Category4'))%>%
inner_join(tibble(value = 'Y',
key = c('Category3', 'Category4'),
lookup = c('Cat3', 'Cat4'))
)%>%
select(Category1, Category2, Category5, Category6 = lookup)%>%
arrange(Category1, Category2)
half_result
> half_result
Category1 Category2 Category5 Category6
1 Phase 1 Apple 5 Cat3
2 Phase 1 Berry 3 Cat4
3 Phase 2 Grape 2 Cat3
4 Phase 2 Grape 2 Cat4
5 Phase 2 Orange 4 Cat3
6 Phase 2 Orange 4 Cat4
half_result%>%
group_by(Category1, Category2)%>%
mutate(category3 = ifelse(any(Category6 == 'Cat3'), 'Y', 'N'),
Category4 = ifelse(any(Category6 == 'Cat4'), 'Y', 'N'))%>%
ungroup()
# A tibble: 6 x 6
Category1 Category2 Category5 Category6 category3 Category4
<fct> <fct> <fct> <chr> <chr> <chr>
1 Phase 1 Apple 5 Cat3 Y N
2 Phase 1 Berry 3 Cat4 N Y
3 Phase 2 Grape 2 Cat3 Y Y
4 Phase 2 Grape 2 Cat4 Y Y
5 Phase 2 Orange 4 Cat3 Y Y
6 Phase 2 Orange 4 Cat4 Y Y