dplyr按字符串分组匹配

我们能做到

library(dplyr)
df %>% 
  group_by(grp = gsub("\\D", "", v1)) %>% 
  summarise(mean_v2 = mean(v2)) # only for illustration
# A tibble: 4 x 2
#  grp   mean_v2
#  <chr>   <dbl>
#1 1      0.0635
#2 2      0.0338
#3 3     -0.130 
#4 4      1.60

这个 group_by 和 gsub v1 ,请参见 ?regex .

set.seed(1)
df <- data.frame(v1 = c('jan1', 'jan2', 'jan3', 'jan4', 'feb1', 'feb2', 'mar1', 'mar2', 'mar3'),
                 v2 = rnorm(9))

str_remove

library(tidyverse)
df %>% 
  group_by(grp = str_remove(v1, "\\D+")) %>% 
  summarise(mean_v2 = mean(v2))
# A tibble: 4 x 2
#  grp   mean_v2
#  <chr>   <dbl>
#1 1      0.0635
#2 2      0.0338
#3 3     -0.130 
#4 4      1.60