CODEWITHSUNDEEP
r
rishi
rishi

Reputation: 267

Making non-unique row names unique

I have a data.frame object in R , for which I want to have non-unique row names:

38.40000 41.75200 44.38400 44.18400 45.37600 37.49600 41.36800 33.93600 38.00800  42.51200 46.49600  40.48000  45.40800  46.32800  43.78400  39.88800  38.84000  40.56800  42.03200 38.89185
45.53846 50.08462 39.91538 36.95385 34.96154 39.74615 38.01538 35.75385 35.54615  36.69231 35.20769  38.05385  39.29231  37.96923  37.30000  36.86923  39.19231  38.81538  43.69231 38.06400 
46.05176 41.69412 38.80000 37.75529 39.67529 39.07765 39.17647 38.24941 39.58588  38.63529 38.30588  41.87765  38.97412  40.13647  42.27294  38.24471  35.41647  40.80000  38.07059 42.11294
44.20000 43.42857 44.80000 35.20000 35.91429 37.82857 51.45714 44.68571 46.68571  48.74286 41.25091  39.45455  38.17091  40.70182  40.39273  41.28727  40.63636  41.50909  41.68364 41.29455
45.06909 41.09818 40.02909 42.50182 42.34909 39.84727 41.42909 40.47273 40.28000  40.51636 41.25091  39.45455  38.17091  40.70182  40.39273  41.28727  40.63636  41.50909  41.68364 41.29455
40.87407 39.27704 44.13630 43.25037 35.86667 37.30667 38.76148 40.74667 38.93333  43.16148 37.47259  37.73630  38.34370  39.00148  36.96889  37.76593  39.14667  37.92593  37.62963 38.89185

The rownames I want for this dummy data would be B,C,C,B,E2,E3. However, I am aware that R does not allow non-unique rownames. I have seven possible rowname classes for my complete dataset, A,B,C,D,E1,E2,E3.

I was trying to write a script in R (but failing), that takes a vector of my non-unique rownames, and adds numbers to the elements 1,2,3... so on, depending on the length of that character in the vector.

Thus for this dummy data, the vector I would have is B-1,C-1,C-2,B-2,E2-1,E3-1, and my final matrix would be:

B-1  38.40000 41.75200 44.38400 44.18400 45.37600 37.49600 41.36800 33.93600 38.00800  42.51200 46.49600  40.48000  45.40800  46.32800  43.78400  39.88800  38.84000  40.56800  42.03200 38.89185
C-1  45.53846 50.08462 39.91538 36.95385 34.96154 39.74615 38.01538 35.75385 35.54615  36.69231 35.20769  38.05385  39.29231  37.96923  37.30000  36.86923  39.19231  38.81538  43.69231 38.06400 
C-2  46.05176 41.69412 38.80000 37.75529 39.67529 39.07765 39.17647 38.24941 39.58588  38.63529 38.30588  41.87765  38.97412  40.13647  42.27294  38.24471  35.41647  40.80000  38.07059 42.11294
B-2  44.20000 43.42857 44.80000 35.20000 35.91429 37.82857 51.45714 44.68571 46.68571  48.74286 41.25091  39.45455  38.17091  40.70182  40.39273  41.28727  40.63636  41.50909  41.68364 41.29455
E2-1 45.06909 41.09818 40.02909 42.50182 42.34909 39.84727 41.42909 40.47273 40.28000  40.51636 41.25091  39.45455  38.17091  40.70182  40.39273  41.28727  40.63636  41.50909  41.68364 41.29455
E3-1 40.87407 39.27704 44.13630 43.25037 35.86667 37.30667 38.76148 40.74667 38.93333  43.16148 37.47259  37.73630  38.34370  39.00148  36.96889  37.76593  39.14667  37.92593  37.62963 38.89185

Upvotes: 2

Views: 836

Answers (3)

pogibas
pogibas

Reputation: 28379

Since this question is how to:

take a vector of non-unique rownames and add numbers to the elements 1,2,3

I will ignore the PCA part (recommendations to use matrix or not to use rownames at all).

To generate wanted vector of names you can use this:

foo <- c("A", "B", "C", "C", "B", "E", "E")
paste0(foo, "-", sapply(seq_along(foo), function(x) sum(foo[1:x] == foo[x])))
[1] "A-1" "B-1" "C-1" "C-2" "B-2" "E-1" "E-2"

Upvotes: 1

Prem
Prem

Reputation: 11985

You can try this. 

df <- data.frame(row_name = c('B','C','C','B','E2','E3'))

library(dplyr)

df <- df %>% 
  group_by(row_name) %>% 
  mutate(count = sequence(n()),
         unique_row_name = paste(row_name, count, sep="-"))

df$unique_row_name is your candidate!

Upvotes: 0

Thomas Mailund
Thomas Mailund

Reputation: 1826

If you really want to do this, then this will work

uniqify_names <- function(names_vector) {
    names <- unique(names_vector)
    count_table <- rep(0, length(names))
    names(count_table) <- names # works because R has weird symbol lookup
    update_name <- function(name) {
        new_name <- paste0(name, ".", count_table[name])
        count_table[name] <<- count_table[name] + 1
        new_name
    }
    vapply(names_vector, update_name, FUN.VALUE = "character")
}

It works like this:

> non_unique_names <- c("A", "B", "A", "A", "B", "C", "A", "B", "C")
> uniqify_names(non_unique_names)
    A     B     A     A     B     C     A     B     C 
"A.0" "B.0" "A.1" "A.2" "B.1" "C.0" "A.3" "B.2" "C.1"

You can set the row names using rownames from this vector.

Upvotes: 1

Related Questions