Question

假设列中的每一行（letter_strings）具有由逗号分隔的可变数量的字符串。例如：

letter_strings

abc, def, ghi, jkl 
mno, pqr
stu, vw, xyz

我想在数据框中查找每个字符串：

letter_strings code

abc YES
def NO
ghi MAYBE
jkl SURE
mno PERHAPS
pqr ALWAYS
stu NEVER
vw NOGO
xyz ABSENT

并在附加列中获取以下相应的行

YES, NO, MAYBE, SURE
PERHAPS, ALWAYS
NEVER, NOGO, ABSENT

这在R中是否可行，我真的不知道如何解决这个问题...

提前致谢！

w ^

Answer 1

1）gusbfn

gsubfn与gsub类似，只是它在查找列表的名称中查找与正则表达式（此处定义为"\\w+"，即一系列单词字符）的匹配项，lookup，将目标字符串中的名称替换为lookup中的值。

library(gsubfn)

lookup <- with(DF2, as.list(setNames(code, letter_strings)))
transform(DF1, codes = gsubfn("\\w+", lookup, letter_strings))

，并提供：

      letter_strings                codes
1 abc, def, ghi, jkl YES, NO, MAYBE, SURE
2           mno, pqr      PERHAPS, ALWAYS
3       stu, vw, xyz  NEVER, NOGO, ABSENT

2）dplyr / tidyr 将DF1转换为长格式，将其与DF2连接，然后将其重新整形为原始格式：

library(dplyr)
library(tidyr)

DF1 %>% 
    mutate(id = 1:n()) %>% 
    separate_rows(letter_strings) %>% 
    left_join(DF2) %>% 
    group_by(id) %>% 
    summarise(letter_string = toString(letter_strings), codes = toString(code)) %>% 
    ungroup %>%
    select(-id)

，并提供：

Joining, by = "letter_strings"
# A tibble: 3 x 2
       letter_string                codes
               <chr>                <chr>
1 abc, def, ghi, jkl YES, NO, MAYBE, SURE
2           mno, pqr      PERHAPS, ALWAYS
3       stu, vw, xyz  NEVER, NOGO, ABSENT

3）strsplit / merge / aggregate 使用strsplit将DF1和stack中的字符串拆分为长格式st。然后merge DF2和aggregate返回原始表单。没有包使用。

s <- strsplit(DF1$letter_strings, ", ")
st <- stack(setNames(s, seq_along(s)))
m <- merge(st, DF2, by = 1, all.x = TRUE, all.y = FALSE)
aggregate(. ~ ind, m, toString)[-1]

，并提供：

              values                 code
1 abc, def, ghi, jkl YES, NO, MAYBE, SURE
2           mno, pqr      PERHAPS, ALWAYS
3       stu, vw, xyz  NEVER, NOGO, ABSENT

3a）magrittr 这可以用magrittr表示：

library(magrittr)

DF1 %>%
    "$"("letter_strings") %>%
    strsplit(", ") %>%
    setNames(seq_along(.)) %>%
    stack %>%
    merge(DF2, by = 1, all.x = TRUE, all.y = FALSE) %>%
    aggregate(. ~ ind, ., toString) %>%
    "["(-1)


s <- stack(setNames(strsplit(DF1$letter_strings, ", "), 1:nrow(DF1)))
m <- merge(s, DF2, by = 1, all.x = TRUE, all.y = FALSE)
aggregate(. ~ ind, m, toString)[-1]

4）data.table 请注意，在下面的评论中，@ Uwe提供了（2）和（3）中转换为长格式的方法的data.table版本，加入并转换回来。

注意：以可重复的形式输入：

Lines1 <- "
letter_strings
abc, def, ghi, jkl
mno, pqr
stu, vw, xyz"
DF1 <- read.table(text = Lines1, header = TRUE, as.is = TRUE, sep = ";")

Lines2 <- "
letter_strings code
abc YES
def NO
ghi MAYBE
jkl SURE
mno PERHAPS
pqr ALWAYS
stu NEVER
vw NOGO
xyz ABSENT"
DF2 <- read.table(text = Lines2, header = TRUE, as.is = TRUE)

Answer 2

如果没有太多字母字符串，您可以在循环中使用gsub执行此操作。

Temp = letter_strings
for(i in 1:nrow(df)) {
    Temp = gsub(df$letter_strings[i], df$code[i], Temp) }
Temp
[1] "YES, NO, MAYBE, SURE" "PERHAPS, ALWAYS"      "NEVER, NOGO, ABSENT"

R中的数据框：查找单元格中由逗号分隔的多个字符串

2 个答案: