R：从混合数字和字母的字符串中提取最大数字

Question

最好，我正在寻找dplyr 解决方案。

我有

> str(p)
'data.frame':   25 obs. of  1 variable:
 $ intram_size: chr  "5" "4,7 x 6,6 mm" "4x6x7 mm" "5" ...

有

> head(p)
   intram_size
1            5
2 4,7 x 6,6 mm
3     4x6x7 mm
4            5
5         4x11
6          1x4

p$intram_size 表示某个肿瘤的二维测量值。我需要提取最大的数字，即测量的最大直径。一个问题是, 已被使用。

Expected output

> head(p)
   intram_size       new
1            5         5
2 4,7 x 6,6 mm       6.6
3     4x6x7 mm         7 
4            5         5
5         4x11        11 
6          1x4         4

数据样本

p <- structure(list(intram_size = c("5", "4,7 x 6,6 mm", "4x6x7 mm", 
"5", "4x11", "1x4", "7x10", "8", "3", "7", "7x4x3", "10x5", "8", 
"7", "11", "7", "10", "5", "13", "5", "3,5", "10", "2,5", "7", 
"11 x 6 x 4")), row.names = c(NA, 25L), class = "data.frame")

Answer 1

1. 用点替换逗号
2. 从字符串中提取所有数字。
3. 转换为数值并返回最大值。

library(tidyverse)

p %>%
  mutate(intram_size = str_replace_all(intram_size, ',', '.'), 
         new = str_extract_all(intram_size, '\\d+(\\.\\d+)?'), 
         new = map_dbl(new, ~max(as.numeric(.x))))

#    intram_size  new
#1             5  5.0
#2  4.7 x 6.6 mm  6.6
#3      4x6x7 mm  7.0
#4             5  5.0
#5          4x11 11.0
#6           1x4  4.0
#7          7x10 10.0
#8             8  8.0
#9             3  3.0
#10            7  7.0
#11        7x4x3  7.0
#12         10x5 10.0
#13            8  8.0
#14            7  7.0
#15           11 11.0
#16            7  7.0
#17           10 10.0
#18            5  5.0
#19           13 13.0
#20            5  5.0
#21          3.5  3.5
#22           10 10.0
#23          2.5  2.5
#24            7  7.0
#25   11 x 6 x 4 11.0

Answer 2

使用 dplyr（添加和修改列）和 stringr（提取模式），过程可能如下所示：

# sample data
p <- structure(list(intram_size = c("5", "4,7 x 6,6 mm", "4x6x7 mm", 
                                    "5", "4x11", "1x4", "7x10", "8", "3", "7", "7x4x3", "10x5", "8", 
                                    "7", "11", "7", "10", "5", "13", "5", "3,5", "10", "2,5", "7", 
                                    "11 x 6 x 4")), row.names = c(NA, 25L), class = "data.frame")
library(dplyr)
library(stringr)
mod <- p %>% 
  # replace decimal separator
  mutate(intram_size = str_replace_all(intram_size, ",", "."),
         # extract numbers
         split = str_extract_all(intram_size, "[0-9\\.]+")) %>% 
  rowwise() %>% 
  # convert to right data type
  mutate(num = list(as.numeric(split)),
         # find maximum
         max = max(num, na.rm = TRUE))

head(mod)
#> # A tibble: 6 x 4
#> # Rowwise: 
#>   intram_size  split     num         max
#>   <chr>        <list>    <list>    <dbl>
#> 1 5            <chr [1]> <dbl [1]>   5  
#> 2 4.7 x 6.6 mm <chr [2]> <dbl [2]>   6.6
#> 3 4x6x7 mm     <chr [3]> <dbl [3]>   7  
#> 4 5            <chr [1]> <dbl [1]>   5  
#> 5 4x11         <chr [2]> <dbl [2]>  11  
#> 6 1x4          <chr [2]> <dbl [2]>   4

^{由reprex package (v0.3.0) 于 2020 年 12 月 3 日创建}