r - Label as the same if the first part of string is the same -

some example data:

    id_trial          001_a.txt            001_a_t2.txt         949482_b.txt         949482_b_t2.txt      95_c.txt             95_c_t2.txt         

note: strings of varying length, length equal pairs minus "_t2"

how can make if part of string before _t2 same both labeled, in new column, such. is, want like:

    id_trial         subject  001_a.txt           person_a  001_a_t2.txt        person_a  949482_b.txt        person_b  949482_b_t2.txt     person_b  95_c.txt            person_c  95_c_t2.txt         person_c 

even work:

    id_trial         subject  001_a.txt                001_a_t2.txt             949482_b.txt            b  949482_b_t2.txt         b  95_c.txt                c  95_c_t2.txt             c 

any appreciated.

you can try sub extract prefix part

df1$subject <-   sub('([^_]+_.).*', '\\1',sub('([^_]+)\\1+',           '\\1', df1$id_trial)) df1 #        id_trial  subject #1   personn_a.txt person_a #2 person_a_t2.txt person_a #3    person_b.txt person_b #4 person_b_t2.txt person_b #5  personnn_c.txt person_c #6 person_c_t2.txt person_c 

if need numeric subject

as.numeric(factor(df1$subject)) #[1] 1 1 2 2 3 3 

update

for second dataset

df2$subject <- sub('\\d+_([a-z]+).*', '\\1', df2$id_trial) df2 #         id_trial subject #1       001_a.txt       #2    001_a_t2.txt       #3    949482_b.txt       b #4 949482_b_t2.txt       b #5        95_c.txt       c #6     95_c_t2.txt       c 

data

df1 <-  structure(list(id_trial = c("personn_a.txt", "person_a_t2.txt",  "person_b.txt", "person_b_t2.txt", "personnn_c.txt", "person_c_t2.txt" )), .names = "id_trial", class = "data.frame", row.names = c(na, -6l))  df2 <- structure(list(id_trial = c("001_a.txt", "001_a_t2.txt",  "949482_b.txt",  "949482_b_t2.txt", "95_c.txt", "95_c_t2.txt")), .names = "id_trial",  class = "data.frame", row.names = c(na, -6l))