CODEWITHSUNDEEP
rranking
Fernanda
Fernanda

Reputation: 313

Adding multiple ranking columns in a dataframe in R

I have the following df:

id city uf home money work
34  LA  RJ  10     2     2   
33  BA  TY  7      3     65
32  NY  BN  4      5     4
12  SP  SD  3      9     7
14  FR  DE  1      8     9
17  BL  DE  5      10    8

DESIRED DF:

id city uf home  rank_home  money    rank_money   work   rank_work
34  LA  RJ  10    1          2         6           2       6
33  BA  TY  7     2          3         5           65      1
32  NY  BN  4     4          5         4           4       5
12  SP  SD  3     5          9         2           7       4
14  FR  DE  1     6          8         3           9       2
17  BL  DE  5     3          10        1           8       3

I know this is possible: dat$rank_home <- rank(dat$home)

But I want a cleaner code for multiple columns!

Thank you!!

Upvotes: 3

Views: 253

Answers (3)

Peace Wang
Peace Wang

Reputation: 2419

One optional method in data.table

library(data.table)
new <- paste0("rank_",names(dt[,4:6]))
dt[,(new):=lapply(-.SD,frankv),.SDcols=4:6]

Upvotes: 1

ThomasIsCoding
ThomasIsCoding

Reputation: 101209

A base R option

cbind(
    df,
    setNames(
        lapply(-df[-(1:3)], rank),
        paste0("rank_", names(df)[-(1:3)])
    )
)

gives

  id city uf home money work rank_home rank_money rank_work
1 34   LA RJ   10     2    2         1          6         6
2 33   BA TY    7     3   65         2          5         1
3 32   NY BN    4     5    4         4          4         5
4 12   SP SD    3     9    7         5          2         4
5 14   FR DE    1     8    9         6          3         2
6 17   BL DE    5    10    8         3          1         3

Upvotes: 3

akrun
akrun

Reputation: 887038

We can loop across the columns 'home' to 'work', apply the rank, while creating new column by adding prefix in .names, and probably select to keep the order

library(dplyr)
df1 <- df %>% 
   mutate(across(home:work, ~ rank(-.), .names = "rank_{.col}"))

Or may do this in a loop where it is more flexible in placing the column at a particular position by specifying either .after or .before. Note that we used compound assignment operator (%<>% from magrittr) to do the assignment in place

library(magrittr)
library(stringr)
for(nm in names(df)[4:6]) df %<>%
     mutate(!!str_c("rank_", nm) := rank(-.data[[nm]]), .after = all_of(nm))

-output

df
  id city uf home rank_home money rank_money work rank_work
1 34   LA RJ   10         1     2          6    2         6
2 33   BA TY    7         2     3          5   65         1
3 32   NY BN    4         4     5          4    4         5
4 12   SP SD    3         5     9          2    7         4
5 14   FR DE    1         6     8          3    9         2
6 17   BL DE    5         3    10          1    8         3

NOTE: If the column have ties, then the default method use is "average". So, ties.method can also be an argument in the rank where there are ties.

data

df <- structure(list(id = c(34L, 33L, 32L, 12L, 14L, 17L), city = c("LA", 
"BA", "NY", "SP", "FR", "BL"), uf = c("RJ", "TY", "BN", "SD", 
"DE", "DE"), home = c(10L, 7L, 4L, 3L, 1L, 5L), money = c(2L, 
3L, 5L, 9L, 8L, 10L), work = c(2L, 65L, 4L, 7L, 9L, 8L)), 
class = "data.frame", row.names = c(NA, 
-6L))

Upvotes: 3

Related Questions