R data frame rank by groups (group by rank) with package dplyr

Question

I have a data frame \'test\' that look like this:

    session_id  seller_feedback_score
1   1   282470
2   1   275258
3   1   275258
4   1   275258
5   1   3         


        

Answer 1

One way you can do this is :

dataset<-dataset%>%arrange(ID, DateTime,Index)
dataset$Rank<-c(0,ID)[-(nrow(dataset)+1)] == ID
dataset<- dataset%>%group_by(ID)%>%mutate(Rank = cumsum(Rank))

Had the same issue!

Answer 2

Had a similar issue, my answer was sorting on groups and the relevant ranked variable(s) in order to then use row_number() when using group_by.

# Sample dataset
df <- data.frame(group=rep(c("GROUP 1", "GROUP 2"),10),
               value=as.integer(rnorm(20, mean=1000, sd=500)))
require(dplyr)
print.data.frame(df[0:10,])
   group value
1  GROUP 1  1273
2  GROUP 2  1261
3  GROUP 1  1189
4  GROUP 2  1390
5  GROUP 1  1942
6  GROUP 2  1111
7  GROUP 1   530
8  GROUP 2   893
9  GROUP 1   997
10 GROUP 2   237

sorted <- df %>% 
          arrange(group, -value) %>%
          group_by(group) %>%
          mutate(rank=row_number())
print.data.frame(sorted)

      group value rank
1  GROUP 1  1942    1
2  GROUP 1  1368    2
3  GROUP 1  1273    3
4  GROUP 1  1249    4
5  GROUP 1  1189    5
6  GROUP 1   997    6
7  GROUP 1   562    7
8  GROUP 1   535    8
9  GROUP 1   530    9
10 GROUP 1     1   10
11 GROUP 2  1472    1
12 GROUP 2  1390    2
13 GROUP 2  1281    3
14 GROUP 2  1261    4
15 GROUP 2  1111    5
16 GROUP 2   893    6
17 GROUP 2   774    7
18 GROUP 2   669    8
19 GROUP 2   631    9
20 GROUP 2   237   10

Answer 3

Found an answer in : Add a "rank" column to a data frame

data.selected <- transform(data.selected, 
              seller_feedback_score_rank = ave(seller_feedback_score, session_id, 
                              FUN = function(x) rank(-x, ties.method = "first")))