dplyr：从子字符串的多个 areas/characters 变异

Question

初学者。尝试使用 dplyr: mutate/case_when 具有多个子字符串条件（10 个字符的字符串）。每个字符代表一个种族。例如，字符串的 1-3 位中的任何一个为“Y”而其余位置为“N”的字符串将被定义为“Latino”。我正在尝试获取一个字符串的正确代码，该字符串在 1-3 个位置中的任何一个位置都有一个“Y”，但在 4-5 个位置（“亚洲”）中的任何一个位置都有一个“Y”。我想将其定义为“多种族”。希望提供正确的代码来创建结果为“多民族”的字符串？非常感谢这个网站！

library(dplyr)

data = data.frame(APP_AC = c("YNNNNNNNNN",
                             "YYNNNNNNNN",
                             "YYYNNNNNNN",
                             "YNYNNNNNNN",
                             "NNNYNNNNNN",
                             "YNNYNNNNNN",
                             "NNNNNYNNNN",
                             "YNNNNYNNNY",
                             "NNNNNNNNNN"))

data %>% 
  mutate(ETHNICITY = case_when(
    str_sub(APP_AC,1,1) == "Y" ~ "Latino", 
    str_sub(APP_AC,2,2) == "Y" ~ "Latino",
    str_sub(APP_AC,3,3) == "Y" ~ "Latino",
    str_sub(APP_AC,4,4) == "Y" ~ "Asian",
    str_sub(APP_AC,5,5) == "Y" ~ "Asian",
    str_sub(APP_AC,6,6) == "Y" ~ "Black",
    str_sub(APP_AC,7,7) == "Y" ~ "Native_American_Alaskan",
    str_sub(APP_AC,8,8) == "Y" ~ "Pacific_Islander",
    str_sub(APP_AC,9,9) == "Y" ~ "Pacific_Islander",
    str_sub(APP_AC,10,10) == "Y" ~ "White",
    TRUE ~ "Unknown"))

    APP_AC     ETHNICITY
1   YNNNNNNNNN Latino
2   YYNNNNNNNN Latino
3   YYYNNNNNNN Latino
4   YNYNNNNNNN Latino
5   NNNYNNNNNN Asian
6   YNNYNNNNNN Asian
7   NNNNNYNNNN Black
8   YNNNNYNNNY Latino
9   NNNNNNNNNN Unknown

期望的输出：

    APP_AC     ETHNICITY
1   YNNNNNNNNN Latino
2   YYNNNNNNNN Latino
3   YYYNNNNNNN Latino
4   YNYNNNNNNN Latino
5   NNNYNNNNNN Asian
6   YNNYNNNNNN Multi-Ethnic
7   NNNNNYNNNN Black
8   YNNNNYNNNY Multi-Ethnic
9   NNNNNNNNNN Unknown

Answer 1

您可以使用 str_detect :

library(dplyr)
library(stringr)

data %>% 
  mutate(ETHNICITY = case_when(
    str_count(APP_AC, 'Y') > 1 ~ "Multi-Ethnic",
    str_sub(APP_AC,1,1) == "Y" ~ "Latino", 
    str_sub(APP_AC,2,2) == "Y" ~ "Latino",
    str_sub(APP_AC,3,3) == "Y" ~ "Latino",
    str_sub(APP_AC,4,4) == "Y" ~ "Asian",
    str_sub(APP_AC,5,5) == "Y" ~ "Asian",
    str_sub(APP_AC,6,6) == "Y" ~ "Black",
    str_sub(APP_AC,7,7) == "Y" ~ "Native_American_Alaskan",
    str_sub(APP_AC,8,8) == "Y" ~ "Pacific_Islander",
    str_sub(APP_AC,9,9) == "Y" ~ "Pacific_Islander",
    str_sub(APP_AC,10,10) == "Y" ~ "White",
    TRUE ~ "Unknown"))

#      APP_AC    ETHNICITY
#1 YNNNNNNNNN       Latino
#2 NNNYNNNNNN        Asian
#3 YNNYNNNNNN Multi-Ethnic
#4 NNNNNYNNNN        Black
#5 NNNNNNNNNN      Unknown

同样，您也可以将其他条件合二为一，使代码更短。

data %>% 
  mutate(ETHNICITY = case_when(
    str_count(APP_AC, 'Y') > 1 ~ "Multi-Ethnic",
    str_detect(str_sub(APP_AC, 1, 3), 'Y') ~ "Latino",
    str_detect(str_sub(APP_AC, 4, 5), 'Y') ~ "Asian",
    str_sub(APP_AC,6,6) == "Y" ~ "Black",
    str_sub(APP_AC,7,7) == "Y" ~ "Native_American_Alaskan",
    str_detect(str_sub(APP_AC, 8, 9), 'Y') ~ "Pacific_Islander",
    str_sub(APP_AC,10,10) == "Y" ~ "White",
    TRUE ~ "Unknown"))

Answer 2

已更新以包含评论中描述的逻辑和更新的问题。

代码

data %>% 
  mutate(ETHNICITY = case_when(
    str_detect(substr(APP_AC, 1, 3),"Y") & str_count(substr(APP_AC, 4, 10), "Y") == 0 ~ "Latino", 
    str_detect(substr(APP_AC, 1, 3),"Y") & str_count(substr(APP_AC, 4, 10), "Y") >= 1 ~ "Multi-Ethnic", 
    str_detect(substr(APP_AC, 4, 5),"Y")  ~ "Asian", 
    str_detect(substr(APP_AC, 6, 6),"Y") ~ "Black",
    str_detect(substr(APP_AC, 7, 7),"Y") ~ "Native_American_Alaskan",
    str_detect(substr(APP_AC, 8, 9),"Y") ~ "Pacific_Islander",
    str_detect(substr(APP_AC, 10, 10),"Y") ~ "White",
    TRUE ~ "Unknown")
  )

输出

      APP_AC    ETHNICITY
1 YNNNNNNNNN       Latino
2 YYNNNNNNNN       Latino
3 YYYNNNNNNN       Latino
4 YNYNNNNNNN       Latino
5 NNNYNNNNNN        Asian
6 YNNYNNNNNN Multi-Ethnic
7 NNNNNYNNNN        Black
8 YNNNNYNNNY Multi-Ethnic
9 NNNNNNNNNN      Unknown

dplyr：从子字符串的多个 areas/characters 变异

dplyr: mutate from multiple areas/characters of a substring

r

dplyr

substr

case-when