根据同一行中的其他值选择值

Question

I have a dataframe (df) in where I have the colums date, winner, loser, WinnerRank and Loserrank. The winnerrank is the rank of the person from the column winner, and the same goes for loserrank. I want to have a new dataframe with the date, name and rank. But the problem is that the name I want can be in both colums, winner and Loser. If the name I want is in the winner column I would like to have the Winnerrank, but if the name is in the Loser I want to have the loserrank. How do I do this?

the df looks like this:

        Date       Winner          Loser WRank LRank
1 2000-01-03   Federer R. Knippschild J.    65    87
2 2000-01-03   Enqvist T.     Federer R.     5    65
3 2000-01-10 Ferrero J.C.     Federer R.    45    61
4 2000-01-17   Federer R.       Chang M.    62    38
5 2000-01-17   Federer R.     Kroslak J.    62   104
6 2000-01-17   Clement A.     Federer R.    54    62

And the format I want looks like this:

        Date       Name    Rank 
1 2000-01-03   Federer R.  65
2 2000-01-03   Federer R.  65   
3 2000-01-10   Federer R.  61   
4 2000-01-17   Federer R.  62    
5 2000-01-17   Federer R.  62   
6 2000-01-17   Federer R.  62   

Answer 1

We can use the functions found in the tidyverse package:

library(tidyverse)
dat %>%
  # create single winner and loser columns,
  # concatenating name and rank together
  unite(Winner, Winner, WRank, sep = "-") %>%
  unite(Loser, Loser, LRank, sep = "-") %>%
  # pivot to be "tall"
  pivot_longer(cols = c("Winner", "Loser")) %>%
  select(-name) %>%
  # reverse concatentation
  separate(value, into = c("Name", "Rank"), sep = "-")

 #   Date       Name           Rank 
 # 1 2000-01-03 Federer_R.     65   
 # 2 2000-01-03 Knippschild J. 87   
 # 3 2000-01-03 Enqvist_T.     5    
 # 4 2000-01-03 Federer R.     65   
 # 5 2000-01-10 Ferrero_J.C.   45   
 # 6 2000-01-10 Federer R.     61   
 # 7 2000-01-17 Federer_R.     62   
 # 8 2000-01-17 Chang M.       38   
 # 9 2000-01-17 Federer_R.     62   
 #10 2000-01-17 Kroslak J.     104  
 #11 2000-01-17 Clement_A.     54   
 #12 2000-01-17 Federer R.     62   

One thing to note is that this will convert your Rank to a character value. You can reverse that using the as.numeric function.

Answer 2

May be a function helps in extracting those rows and values based on the player name. We filter the rows where the player name is in either 'Winner' or | 'Loser' column, then use transmute to create the three column output by selecting the 'Date', 'Name' as the input player name, and 'Rank' by creating a logical matrix by comparing the subset of columns 'Winner', 'Loser' with player name, feed that output into max.col to get the index of max value ie TRUE => 1 and FALSE => 0 for each row, cbind with row index ( row_number ) and use that to extract the corresponding elements from subset of dataset with 'WRank', 'LRank' columns

f1 <- function(dat, nm) {
      dat %>%
        filter(Winner == nm|Loser == nm) %>%
            transmute(Date, Name = nm, 
        Rank = .[c('WRank', 'LRank')][cbind(row_number(),
             max.col(.[c('Winner', 'Loser')] == nm))])

}

-testing

f1(df1, 'Federer R.')
#        Date       Name Rank
#1 2000-01-03 Federer R.   65
#2 2000-01-03 Federer R.   65
#3 2000-01-10 Federer R.   61
#4 2000-01-17 Federer R.   62
#5 2000-01-17 Federer R.   62
#6 2000-01-17 Federer R.   62

data

df1 <- structure(list(Date = c("2000-01-03", "2000-01-03", "2000-01-10", 
"2000-01-17", "2000-01-17", "2000-01-17"), Winner = c("Federer R.", 
"Enqvist T.", "Ferrero J.C.", "Federer R.", "Federer R.", "Clement A."
), Loser = c("Knippschild J.", "Federer R.", "Federer R.", "Chang M.", 
"Kroslak J.", "Federer R."), WRank = c(65L, 5L, 45L, 62L, 62L, 
54L), LRank = c(87L, 65L, 61L, 38L, 104L, 62L)),
class = "data.frame", row.names = c("1", 
"2", "3", "4", "5", "6"))

Answer 3

You can combine both winner and loser in one column and rank into one column and then select the player name.

library(dplyr)

player_name <- 'Federer R.'

df %>%
  rename_with(~paste0(., '_Name'), c(Winner, Loser)) %>%
  rename_with(~paste0(., '_Rank'), ends_with('Rank')) %>%
  tidyr::pivot_longer(cols = -Date, 
               names_pattern = '.*_(\\w+)', 
               names_to = '.value') %>%
  filter(Name == player_name)

#   Date       Name        Rank
#  <chr>      <chr>      <int>
#1 2000-01-03 Federer R.    65
#2 2000-01-03 Federer R.    65
#3 2000-01-10 Federer R.    61
#4 2000-01-17 Federer R.    62
#5 2000-01-17 Federer R.    62
#6 2000-01-17 Federer R.    62