Ytukyg

I have a data table that contains character observations:

library(data.table)

library(stringr)



DT = data.table(strings = c('AAABD', 'BBDA', 'AACBDAA', 'ABACD'))

I would like to create a variable that contains counts of 'A', 'AA', and 'AAA' in each observation as a list. To do this I have created a function foo:

foo <- function(str) {

    n <- str_count(str, 'A')

    n2 <- str_count(str, 'AA')

    n3 <- str_count(str, 'AAA')

    df <- list('n' = n, 'n2' = n2, 'n3' = n3)



    return(df)

}

I apply this function to DT to create a new variable for count observation as a list:

DT[, count := foo(strings)]

When I do this, I receive this error:

Warning message:

In `[.data.table`(DT, , `:=`(counts, foo(strings))) :

  Supplied 3 items to be assigned to 4 items of column 'counts' (recycled leaving remainder of 1 items).

The data table returned has count variable lists of size 4 instead of size 3 and does not represent the amount of 'A', 'AA', and 'AAA' accurately for each string observation in variable strings. How can I assign a list as an observation in a data table?

You need to transpose the list:

foo <- function(str) {

  n <- str_count(str, 'A')

  n2 <- str_count(str, 'AA')

  n3 <- str_count(str, 'AAA')

  df <- transpose(list('n' = n, 'n2' = n2, 'n3' = n3)) # <- add transpose 

  return(df)

}



DT[, count := foo(strings)]

DT

#     strings count

#  1:   AAABD 3,1,1

#  2:    BBDA 1,0,0

#  3: AACBDAA 4,2,0

#  4:   ABACD 2,0,0



str(DT)

# Classes ‘data.table’ and 'data.frame':    4 obs. of  2 variables:

# $ strings: chr  "AAABD" "BBDA" "AACBDAA" "ABACD"

# $ count  :List of 4

#  ..$ : int  3 1 1

#  ..$ : int  1 0 0

#  ..$ : int  4 2 0

#  ..$ : int  2 0 0

Assigning a list of columns to a single name can result in the warning message. Instead it can be

DT[, c('n', 'n2', 'n3') := .(str_count(strings, 'A'), 

        str_count(strings, 'AA'),  str_count(strings, 'AAA'))]