Is there a R function to strsplit specifc rows under one column based on another column?

Question

I am having a data frame (for example as below:

name  student_id   age  gender
Sam   123_abc_ABC  20   F
John  234_bcd_BCD  18   M
Mark  345_cde_CDE  20   M
Ram   xyz_111_XYZ  19   M
Hari  uvw_444_UVW  23   M

Now, I need a new column as student_id_by_govt in the df. The student_id_by_govt is within the student_id but it is different for different names. For Sam, John, Mark the student_id_by_govt would be first segment of student_id (i.e., 123, 234, 345) but for Ram & Hari, the student_id_by_govt is second segment in the student_id (i.e.,111, 444).

I used the strsplit, lapply commands to get the specfic segment from the student_id but I could not able to apply that command specifically for specific rows to get the desired output mentioned above. Please let me know how to get the output as below:

name  student_id   age  gender student_id_by_govt
Sam   123_abc_ABC  20   F      123
John  234_bcd_BCD  18   M      234
Mark  345_cde_CDE  20   M      345
Ram   xyz_111_XYZ  19   M      111
Hari  uvw_444_UVW  23   M      444

Answer 1

You only need str_extract:

library(tidyverse)
df %>%
  mutate(student_id_by_govt = str_extract(student_id, "\\d+"))
# A tibble: 5 × 3
  Name  student_id  student_id_by_govt
  <chr> <chr>       <chr>             
1 Sam   123_abc_ABC 123               
2 John  234_bcd_BCD 234               
3 Mark  345_cde_CDE 345               
4 Ram   xyz_111_XYZ 111               
5 Hari  uvw_444_UVW 444 

EDIT:

If the student_id_by_govt is determined by gender, as OP notes in comment: "for all the 'M' gender, the student_id_by_govt is first segment in the student_id (ie., 234, 345, xyz, uvw) but for 'F' gender I want the second segment of student_id (i.e., abc)", then this works:

df %>%
  mutate(student_id_by_govt = ifelse(gender == "M", str_extract(student_id, "^[^_]+"),
                                     str_extract(student_id, "(?<=_)[^_]+(?=_)")))
  name  student_id age gender student_id_by_govt
1  Sam 123_abc_ABC  20      F                abc
2 John 234_bcd_BCD  18      M                234
3 Mark 345_cde_CDE  20      F                cde
4  Ram xyz_111_XYZ  19      M                xyz
5 Hari uvw_444_UVW  23      M                uvw

Here, we essentially rely on the negative character class [^_]+, which matches any character but the underscore one or more times as well as positive lookbehind (?<=_)(meaning "match only if there is an underscore to the left") and lookahead (?=_)(meaning "match only if there is an underscore to the right").

Data for solution in EDIT:

df <- read.table(text="name  student_id   age  gender
Sam   123_abc_ABC  20   F
John  234_bcd_BCD  18   M
Mark  345_cde_CDE  20   F
Ram   xyz_111_XYZ  19   M
Hari  uvw_444_UVW  23   M", header = TRUE)