如何从R中的复杂字符串中提取多个新列(str_sub似乎做不到这一点)

Question

我很难从R中的字符串中提取多个变量。

列如下所示：

7 digit identifier_NAME4 digit value-4 digit value-4 digit value-location1-location27-digit identifier_junk_junk_3 digit value with junk attached

例如：

1234567_NAME:0011-1234-0176-town-car:1234567_000001_original_010qyz

我需要有以下内容的新专栏：

• 7位标识符("1234567")
• 名字
• 4位数值中的每一位
• 每一地点
• 3位数值

str_sub()无法工作，因为字符串的某些部分的长度是可变的。

我尝试了gsub，但是由于一些特殊字符被多次重复(即":‘和"-")，所以我无法使用它们提取字符串的一个定义良好的部分。

Answer 1

为了避免冗长的正则表达式，一个选项是使用str_split_fixed作为分隔符，将列拆分成一个矩阵，删除不需要的列，并从最后一列中提取数值：

s <- "1234567_NAME:0011-1234-0176-town-car:1234567_000001_original_010qyz"
ss <- c(s,s,s)

library(stringr)
mat <- str_split_fixed(ss, "[_:-]", 11)[,-c(9, 10)]
mat
#     [,1]      [,2]   [,3]   [,4]   [,5]   [,6]   [,7]  [,8]      [,9]    
#[1,] "1234567" "NAME" "0011" "1234" "0176" "town" "car" "1234567" "010qyz"
#[2,] "1234567" "NAME" "0011" "1234" "0176" "town" "car" "1234567" "010qyz"
#[3,] "1234567" "NAME" "0011" "1234" "0176" "town" "car" "1234567" "010qyz"

mat[,9] <- sub("(\\d{3}).*", "\\1", mat[,9])
mat
#     [,1]      [,2]   [,3]   [,4]   [,5]   [,6]   [,7]  [,8]      [,9] 
#[1,] "1234567" "NAME" "0011" "1234" "0176" "town" "car" "1234567" "010"
#[2,] "1234567" "NAME" "0011" "1234" "0176" "town" "car" "1234567" "010"
#[3,] "1234567" "NAME" "0011" "1234" "0176" "town" "car" "1234567" "010"

Answer 2

如果您保留“冗长”正则表达式，那么您可以这样做(并在后面添加记录验证，因为您已经嵌入了字段长度)：

library(stringi)
library(purrr)

pat <- "(.{7})_([[:alnum:][:space:]]+):([[:digit:]]{4})-([[:digit:]]{4})-([[:digit:]]{4})-([[:alnum:][:space:]]+)-([[:alnum:][:space:]]+):([[:digit:]]{7})_[[:alnum:][:space:]]+_[[:alnum:][:space:]]+_([[:digit:]]{3})"

dat <- "1234567_NAME:0011-1234-0176-town-car:1234567_000001_original_010qyz"
dat <- rep(dat, 10)

cols <- c("id", "name", "val1", "val2", "val3", "loc1", "loc2", "val3")

stri_match_all_regex(dat, pat) %>% 
  map_df(~setNames(as.list(.[,c(2:8,10)]), cols))
## # A tibble: 10 x 7
##         id  name  val1  val2  val3  loc1  loc2
##      <chr> <chr> <chr> <chr> <chr> <chr> <chr>
## 1  1234567  NAME  0011  1234   010  town   car
## 2  1234567  NAME  0011  1234   010  town   car
## 3  1234567  NAME  0011  1234   010  town   car
## 4  1234567  NAME  0011  1234   010  town   car
## 5  1234567  NAME  0011  1234   010  town   car
## 6  1234567  NAME  0011  1234   010  town   car
## 7  1234567  NAME  0011  1234   010  town   car
## 8  1234567  NAME  0011  1234   010  town   car
## 9  1234567  NAME  0011  1234   010  town   car
## 10 1234567  NAME  0011  1234   010  town   car