Categorize according to reference number

[naja]

Hi there:

I have four types of classifications (reference numbers) that I wish to group into a dummy variable. Points 1-3 should be labelled 1 and Point 4 should be labelled 2.

Dummy variable = 1: Internal recruitrment

1. References starting with letters "MH" followed by a five-digit unique identification number (i.e. MH12345, MH45678, MH98743 etc.)

2. References starting with letters "TM" followed by a five-digit unique identification number (i.e. TM12345, TM45678, TM98743 etc.)

3. References that are purely numeric containing seven digits (ie. 1234567, 4657893, 5480238 etc. )

Dummy variable = 2: External recruitment
4) References starting with letters "FB" followed by a five-digit unique identification number (i.e. FB12345, FB45678, FB98743 etc.)

Any ideas on how to do this based on the reference number?
Many Thanks,
Naja

[andresrcs]

You could do something like this

library(tidyverse)

# Sample data
df <- data.frame(stringsAsFactors = FALSE,
                 reference = c("MH12345", "MH45678", "MH98743", "TM12345",
                               "TM45678", "TM98743", "1234567", "4657893",
                               "5480238", "FB12345", "FB45678", "FB98743")
                 )

df %>% 
    mutate(dummy = case_when(
        str_detect(reference, "^(MH|TM|\\d{2})\\d{5}") ~ 1,
        str_detect(reference, "^FB\\d{5}") ~ 2
    ))
#>    reference dummy
#> 1    MH12345     1
#> 2    MH45678     1
#> 3    MH98743     1
#> 4    TM12345     1
#> 5    TM45678     1
#> 6    TM98743     1
#> 7    1234567     1
#> 8    4657893     1
#> 9    5480238     1
#> 10   FB12345     2
#> 11   FB45678     2
#> 12   FB98743     2

# Or if there are no other possible values you could simply do this
df %>% 
    mutate(dummy = if_else(str_detect(reference, "^FB\\d{5}"), 2, 1))
#>    reference dummy
#> 1    MH12345     1
#> 2    MH45678     1
#> 3    MH98743     1
#> 4    TM12345     1
#> 5    TM45678     1
#> 6    TM98743     1
#> 7    1234567     1
#> 8    4657893     1
#> 9    5480238     1
#> 10   FB12345     2
#> 11   FB45678     2
#> 12   FB98743     2

^{Created on 2019-11-07 by the reprex package (v0.3.0.9000)}

[naja]

Hi Andrercs,
How do I then merge this new dummy variable to the dataframe? I get the results for each row as a output when i run the code, but i need the new dummy variable to merge with my existing dataset in order to run tests on it. Is there an easy way to do this?
Many Thanks,
Naja

[andresrcs]

dplyr doesn't perform changes in-place, it creates a new data frame instead, if you want to overwrite the original data frame you have to assign the result explicitly, like this.

df <- df %>% 
    mutate(dummy = if_else(str_detect(reference, "^FB\\d{5}"), 2, 1))

[naja]

Thank you so much! This works now.