Create new variable based on a condition across multiple columns - r

Question

I have a binary variable ("Penalty") and 30 factors with the same levels: "Discharge", "Suspended", "Fine", "Community order", and "Imprisonment".
A small example:
ID
Possession
Importation
Production
Penalty
1
Fine
NA
Fine
Yes
2
NA
NA
Community order
No
3
Discharge
Discharge
NA
No
4
NA
NA
Suspended
Yes
5
Imprisonment
NA
NA
No
6
Fine
NA
Imprisonment
No
I would like to create a new factor based on the same condition across these columns plus the binary variable and where there are differing levels in the same row would like the new variable 'sentence' to retain the levels with this priority: Imprisonment > Community order, Suspended > Fine > Discharge. e.g. Discharge will only be present in the new column where no other level appears.
Desired output:
ID
Possession
Importation
Production
Penalty
Sentence
1
Fine
NA
Fine
Yes
Fine
2
NA
NA
Community order
No
Community order
3
Discharge
Discharge
NA
No
Discharge
4
NA
NA
Suspended
Yes
Suspended
5
Imprisonment
NA
NA
No
Imprisonment
6
Fine
NA
Imprisonment
No
Imprisonment
This is what I have attempted: (where "vec" is a vector of the factor column indices)
data <- data %>%
mutate(
crim_sanct = case_when(
(if_any(vec) == "Discharge") ~ "Discharge",
(if_any(vec) == "Fine") | (data$Penalty == "Yes") ~ "Fine",
(if_any(vec) == "Suspended") ~ "Suspended",
(if_any(vec) == "Community order") ~ "Community order",
(if_any(vec) == "Imprisonment") ~ "imprisonment"))

Answer 1

You are in the right direction but have some small syntax issues in if_any.
Also in case_when you need to put the conditions based on the priority. So if Imprisonment > Community order then Imprisonment condition should come first before Community order.
library(dplyr)
data <- data %>%
mutate(
crim_sanct =
case_when(
if_any(Possession:Production, ~. == "Imprisonment") ~ "imprisonment",
if_any(Possession:Production, ~ . == "Discharge") ~ "Discharge",
if_any(Possession:Production, ~. == "Suspended") ~ "Suspended",
if_any(Possession:Production, ~. == "Fine") | (Penalty == "Yes") ~ "Fine",
if_any(Possession:Production, ~. == "Community order") ~ "Community order")
)
data
# ID Possession Importation Production Penalty crim_sanct
#1 1 Fine <NA> Fine Yes Fine
#2 2 <NA> <NA> Community order No Community order
#3 3 Discharge Discharge <NA> No Discharge
#4 4 <NA> <NA> Suspended Yes Suspended
#5 5 Imprisonment <NA> <NA> No imprisonment
#6 6 Fine <NA> Imprisonment No imprisonment

Answer 2

Since I don't know how to handle the Penalty column, we ignore it for now. Creating a column Sentence based on the columns Possession, Importation and Production could be done with
library(dplyr)
data %>%
mutate(across(
Possession:Production,
~ factor(.x,
c("Imprisonment", "Community order", "Suspended", "Fine", "Discharge"),
ordered = TRUE))) %>%
rowwise() %>%
mutate(Sentence = min(c_across(Possession:Production), na.rm = TRUE)) %>%
ungroup()
which returns
# A tibble: 6 x 6
ID Possession Importation Production Penalty Sentence
<dbl> <ord> <ord> <ord> <chr> <ord>
1 1 Fine NA Fine Yes Fine
2 2 NA NA Community order No Community order
3 3 Discharge Discharge NA No Discharge
4 4 NA NA Suspended Yes Suspended
5 5 Imprisonment NA NA No Imprisonment
6 6 Fine NA Imprisonment No Imprisonment
The main idea here is creating ordered factors and using a rowwise min-function to get the sentence with the hightest priority.
Data
data <- structure(list(ID = c(1, 2, 3, 4, 5, 6), Possession = c("Fine",
NA, "Discharge", NA, "Imprisonment", "Fine"), Importation = c(NA,
NA, "Discharge", NA, NA, NA), Production = c("Fine", "Community order",
NA, "Suspended", NA, "Imprisonment"), Penalty = c("Yes", "No",
"No", "Yes", "No", "No")), problems = structure(list(row = 6L,
col = "Penalty", expected = "", actual = "embedded null",
file = "literal data"), row.names = c(NA, -1L), class = c("tbl_df",
"tbl", "data.frame")), class = "data.frame", row.names = c(NA,
-6L), spec = structure(list(cols = list(ID = structure(list(), class = c("collector_double",
"collector")), Possession = structure(list(), class = c("collector_character",
"collector")), Importation = structure(list(), class = c("collector_character",
"collector")), Production = structure(list(), class = c("collector_character",
"collector")), Penalty = structure(list(), class = c("collector_character",
"collector"))), default = structure(list(), class = c("collector_guess",
"collector")), skip = 1L), class = "col_spec"))