Partial matching string in R-CodePudding

I have a data with 2 variables var and text. I need to remove the partial matching from text and save it. I have the sample of the result.

dt <- data.frame(var  = c("OCILY10","SUDHL16","u2932"), 
                 text = c("OCILY-10-Cas9_T12", "SU-DHL-16_T12_vs_T0","U2932_T10-122-SEMI-A"))

> dt 
      var                 text
1 OCILY10    OCILY-10-Cas9_T12
2 SUDHL16  SU-DHL-16_T12_vs_T0
3   u2932 U2932_T10-122-SEMI-A

#======================
# Result 

dt <- data.frame(var  = c("OCILY10","SUDHL16","u2932"),
                 text = c("OCILY-10-Cas9_T12", "SU-DHL-16_T12_vs_T0","U2932_T10-122-SEMI-A"),
                 result = c("Cas9_T12", "T12_vs_T0","T10-122-SEMI-A"))


> dt
      var                 text         result
1 OCILY10    OCILY-10-Cas9_T12       Cas9_T12
2 SUDHL16  SU-DHL-16_T12_vs_T0      T12_vs_T0
3   u2932 U2932_T10-122-SEMI-A T10-122-SEMI-A

CodePudding user response：

Borrowing from @Akrun's comment, here's a dplyr solution in steps:

library(dplyr)
library(stringr)
dt %>%
  mutate(
    # create temporary column with two first and two last characters from `var`;
    # add `.*` between them as regex metacharacters (@Akrun's ingenious idea!):
    temp = str_replace(var, "^(..).*(..)$", "\\1.*\\2"),
    # make `temp` a case-insensitive pattern und replace what it matches in `text` with nothing:
    result = str_replace(text, paste0("(?i)", temp), ""),
    # remove leading `-` and `_`:
    result = str_replace(result, "^(-|_)", "")) %>%
  # remove temporary column:
  select(-temp)
      var                 text         result
1 OCILY10    OCILY-10-Cas9_T12       Cas9_T12
2 SUDHL16  SU-DHL-16_T12_vs_T0      T12_vs_T0
3   u2932 U2932_T10-122-SEMI-A T10-122-SEMI-A

CodePudding user response：

Another tidyverse solution.

library(tidyverse)

dt %>%
  mutate(split = str_split(text, '(?<=[_-])')) %>%
  unnest(split) %>%
  mutate(sieve = str_remove(split, '_|-')) %>%
  filter(!str_detect(var, str_c('(?i)', sieve))) %>%
  group_by(var) %>%
  mutate(result = str_c(split, collapse = '')) %>%
  distinct(var, text, result)

# # A tibble: 3 x 3
# # Groups:   var [3]
#   var     text                 result        
#   <chr>   <chr>                <chr>         
# 1 OCILY10 OCILY-10-Cas9_T12    Cas9_T12      
# 2 SUDHL16 SU-DHL-16_T12_vs_T0  T12_vs_T0     
# 3 u2932   U2932_T10-122-SEMI-A T10-122-SEMI-A