Combining rows into one row, choosing highest score - r

Question

I have a large dataset with multiple instances of 'Player', and I want to collapse their tournament scores into a single row but keep the row with the highest earnings. It's a large data set so I can't exactly paste the whole thing but a small example table looks like this:
Player
Earned
T1.
T2
T3
John Doe
2100
5
N/A
N/A
John Doe
1900
N/A
12
N/A
John Doe
500
N/A
N/A
16
I'd like to eliminate the N/A values and the duplicate rows by combining Tournament 1, 2, 3 into a single row, and also keeping the highest earnings value (2100), so that it looks more like this:
Player
Earned
T1.
T2
T3
John Doe
2100
5
12
16
So far I've used top_n(1, Earned) to keep the row with the highest earnings, but it only keeps the score of Tournament1, and I need to fill in the other columns with their scores.

Answer 1

From your reference to top_n, I'm inferring dplyr and related packages.
If your columns are strings (since "N/A" is not NA), then
func <- function(x, na.rm = TRUE, na = c("NA", "N/A")) {
if (is.numeric(x)) max(x, na.rm = na.rm) else head(na.omit(setdiff(x, na)), 1)
}
library(dplyr)
dat %>%
group_by(Player) %>%
summarize(across(everything(), func))
# # A tibble: 1 x 5
# Player Earned T1. T2 T3
# <chr> <int> <chr> <chr> <chr>
# 1 John Doe 2100 5 12 16
If your columns are numeric, though, then we can simplify that to
dat %>%
# an interim line to change your strings to numbers
mutate(across(-Player, ~ suppressWarnings(as.numeric(.)))) %>%
# pick up from here
group_by(Player) %>%
summarize(across(everything(), ~ max(., na.rm = TRUE)))
# # A tibble: 1 x 5
# Player Earned T1. T2 T3
# <chr> <dbl> <dbl> <dbl> <dbl>
# 1 John Doe 2100 5 12 16
Data
dat <- structure(list(Player = c("John Doe", "John Doe", "John Doe"), Earned = c(2100L, 1900L, 500L), T1. = c("5", "N/A", "N/A" ), T2 = c("N/A", "12", "N/A"), T3 = c("N/A", "N/A", "16")), class = "data.frame", row.names = c(NA, -3L))

Answer 2

I could not think of any other way to get rid of those N/A values and since you were trying to collapse scores into a single row, I guessed they might be numeric values. Hence I changed them into numeric type:
library(dplyr)
dat %>%
group_by(Player) %>%
mutate(Earned = cummax(Earned),
across(T1:T3, suppressWarnings(as.numeric))) %>%
group_by(Player, Earned) %>%
summarise(across(T1:T3, ~ na.omit(.x)))
# A tibble: 1 x 5
Player Max_Earned T1 T2 T3
<chr> <int> <dbl> <dbl> <dbl>
1 John Doe 2100 5 12 16
I used reproducible data shared by dear #r2evans, so I would like to thank him for that.
(P.S = I changed the T1. to T1 before using it.

Answer 3

Another solution to the top presented solution could be with lead and slice
dat %>%
arrange(desc(Earned)) %>%
mutate(T2 = lead(T2),
T3 = lead(T3,2)) %>%
slice(which.max(Earned))
Output:
Player Earned T1. T2 T3
1 John Doe 2100 5 12 16

Answer 4

Here is a data.table option
> setDT(dat)[, lapply(.SD, function(x) max(x, na.rm = TRUE)), Player]
Player Earned T1. T2 T3
1: John Doe 2100 5 12 16