计算R中多列中的字符值

Question

我有一个这样的Dataframe：

teammember <- c('Member A', 'Member B', 'Member C')
value_a <- c('success', 'fail', NA)
value_b <- c('fail', NA, 'success')
value_c <- c('success', NA, 'fail')

data_df <- data.frame(teammember, value_a, value_b, value_c)

现在我想计算团队成员分组的每个“成功”。我的想法是这样的：

data_df %>%
  group_by(teammember) %>% 
  filter(value_a == "success" | value_b == "success" | value_c == "success") %>% 
  summarise(sales = length(value_a) , length(value_b) , length(value_c)) %>% 
  select(teammember, sales)

我的结果看起来像这样：

# A tibble: 2 x 2
teammember sales
<fct>      <int>
1 Member A       1
2 Member C       1

但它应该是这样的：

# A tibble: 2 x 2
teammember sales
<fct>      <int>
1 Member A       2
2 Member C       1

你能告诉我正确的解决方案应该是什么样的吗？ :)

在此先感谢您的帮助。

康斯坦丁

Answer 1

你用success计算每列的summarize，而你想要的是每行中successs的数量。你可以尝试rowSums：

res <- data.frame(
    teammember = data_df$teammember,
    sales = rowSums(data_df[, paste0('value_', letters[1:3])] == 'success', na.rm = T)
)
#   teammember sales
# 1   Member A     2
# 2   Member B     0
# 3   Member C     1

可以使用res <- res[res$sales > 0, ]删除零值的行。

Answer 2

一种选择是在开始时使用filter_at，然后gather将'value_'列变为'long'格式，filterthe'val有“成功”字符串并获得count

library(dplyr)
library(tidyr)
data_df %>%
    filter_at(vars(matches("value")), any_vars(. %in% 'success')) %>% 
    gather(var, val, value_a:value_c, na.rm = TRUE) %>% 
    filter(val =='success') %>% 
    count(teammember)
# A tibble: 2 x 2
#  teammember     n
#   <fctr>     <int>
#1 Member A       2
#2 Member C       1

---

或者另一个选择是做nesting然后用map我们得到计数

library(purrr)
data_df %>% 
  nest(-teammember) %>%
  transmute(teammember, sales = map(data, ~ sum(unlist(.x) == "success", na.rm = TRUE))) %>% 
  filter(sales != 0)
#   teammember sales
#1   Member A     2
#2   Member C     1