如果你只关心精确匹配：

• 将code和subs转换为如下字符串：

code_string = paste(code, collapse = '')

（类似于subs_string）

• 使用{stringr}查找匹配项：

library(stringr)
str_locate_all(string = 'ABCXABCABC', pattern = 'ABC')

[[1]]
     start end
[1,]     1   3
[2,]     5   7

如果要允许一定量的偏差，您可以：

• 将codestring分割成frame_size长度的连续段（例如，当查找3个字母的模式时，长度为4）：

frame_size = 4
nc = nchar(code_string)

fragments <- substring(code_string,
                       1:(nc - frame_size + 1),
                       frame_size:nc
                       )

> fragments
[1] "ABCD" "BCDE"

• 使用每个片段和你的模式之间的Levenshtein距离（相异性）：

library(comparator)

fragments |> 
  ## `Map` = do this for each item of fragments
  Map(f = \(fragment){ifelse(grepl(subs_string, fragment),
                             ## return 0 for exact match:
                             0,
                             ## penalize insertion or substitution of letters
                             ## in `fragment` for later filtering:
                             Levenshtein(insertion = 100,
                                         deletion = 1,
                                         substitution = 100
                                         )(fragment, subs_string)
                             )
  }
)

输出：

$ABCD
[1] 1 ## this score is OK, we only need to delete one letter from 'ABCD'
      ## to match 'ABD'

$BCDE
[1] 101 ## this score tells us that a letter has been inserted and/or
        ## substituted from 'BCDE' which we don't want

• 现在你可以过滤输出的分数〈2（最多从4个字母的字符串中删除一个以匹配你的3个字母的字符串;这是你的80%相似度）