如何有效地检查R中的特定值并标记包含该值的行中的变量？ 数据

如何解决如何有效地检查R中的特定值并标记包含该值的行中的变量？ 数据

我想创建一个变量，以标记多个变量中的一个或多个是否具有特定值。

week  Mon  Tues  Weds  Thurs  Fri  Sat
1     jon  jon   jon   jon    mary mary
2     jane jane  jane  jane   jane jane
3     mary mary  mary  mary   mary jane

我想创建一个二进制变量，用于每周标记该周的Mon，Wed或Sat ==“ jon”或“ mary”是否有办法做到这一点而无需创建长的ifelse语句来检查每个变量单独变量？

week  Mon  Tues  Weds  Thurs  Fri  Sat  flag
1     jon  jon   jon   jon    mary mary 1
2     jane jane  jane  jane   jane jane 0
3     mary mary  mary  mary   mary jane 1

我尝试过

df %>%
  rowwise() %>%
  mutate(flag = +any(c_across(Mon,Weds,Sat)
  %in% ("jon","mary")) %>%
  ungroup()

但是我得到一个错误

Error: Problem with `mutate()` input `flag`.
x unused arguments (Mon,Sat)
i Input `flag` is `+...`.
i The error occurred in row 1.

解决方法

df %>%
  mutate(flag = colSums(apply(cbind(Mon,Weds,Sat),1,`%in%`,c("jon","mary"))) > 0)
#   week  Mon Tues Weds Thurs  Fri  Sat  flag
# 1    1  jon  jon  jon   jon mary mary  TRUE
# 2    2 jane jane jane  jane jane jane FALSE
# 3    3 mary mary mary  mary mary jane  TRUE

我认为across的问题在于，它试图对每个列进行操作，而不是对所有这些内容进行汇总。让我们尝试purrr::pmap代替L

library(purrr)
df %>%
  mutate(flag = pmap(list(Mon,~ +any(unlist(...) %in% c("jon","mary"))))
#   week  Mon Tues Weds Thurs  Fri  Sat flag
# 1    1  jon  jon  jon   jon mary mary    1
# 2    2 jane jane jane  jane jane jane    0
# 3    3 mary mary mary  mary mary jane    1

三分之一（使用您对c_across的请求）：

df %>%
  rowwise() %>%
  mutate(flag = +any(c_across(c(Mon,Sat)) %in% c("jon","mary"))) %>%
  ungroup()
# # A tibble: 3 x 8
#    week Mon   Tues  Weds  Thurs Fri   Sat    flag
#   <int> <chr> <chr> <chr> <chr> <chr> <chr> <int>
# 1     1 jon   jon   jon   jon   mary  mary      1
# 2     2 jane  jane  jane  jane  jane  jane      0
# 3     3 mary  mary  mary  mary  mary  jane      1

,

如果不使用rowwise和map遍历列，则可以提高效率，而不是reduce遍历

library(purrr)
library(dplyr)
df %>%
     mutate(flag = map(select(.,Mon,"mary")) %>%
          reduce(`|`) %>% `+`)
#  week  Mon Tues Weds Thurs  Fri  Sat flag
#1    1  jon  jon  jon   jon mary mary    1
#2    2 jane jane jane  jane jane jane    0
#3    3 mary mary mary  mary mary jane    1

---

base R中的对应选项是lapply/Reduce

df$flag <- +(Reduce(`|`,lapply(df[c('Mon','Weds','Sat')],"mary"))))

数据

df <- structure(list(week = 1:3,Mon = c("jon","jane","mary"),Tues = c("jon",Weds = c("jon",Thurs = c("jon",Fri = c("mary",Sat = c("mary","jane")),class = "data.frame",row.names = c(NA,-3L
))

,

这是使用rowSums + Reduce

的另一个基本R选项

df$flag <- +(rowSums(
  Reduce(
    `+`,lapply(
      c("jon",`==`,df[c("Mon","Weds","Sat")]
    )
  )
) > 0)

这样

  week  Mon Tues Weds Thurs  Fri  Sat flag
1    1  jon  jon  jon   jon mary mary    1
2    2 jane jane jane  jane jane jane    0
3    3 mary mary mary  mary mary jane    1