How can I avoid a loop when replacing values in a dataframe by row-column indexing?

Question

I want to be able to replace values in a data frame by indexing by row and column, given a list of row indices, column names and values.

library(tidyverse)
cols <- sample(letters[1:10], 5)
vals <- sample.int(5)
rows <- sample.int(5)
df <- matrix(rep(0L, times = 50), ncol = 10) %>%
  `colnames<-`(letters[1:10]) %>%
  as_tibble

I can do this with a for loop over a list of the parameters:

items <- list(cols, vals, rows) %>%
  pmap(~ list(..3, ..1, ..2))

for (i in items){
  df[i[[1]], i[[2]]] <- i[[3]]
}
df
#> # A tibble: 5 x 10
#>       a     b     c     d     e     f     g     h     i     j
#>   <int> <int> <int> <int> <int> <int> <int> <int> <int> <int>
#> 1     0     0     0     0     0     0     0     1     0     0
#> 2     0     0     5     0     0     0     0     0     0     0
#> 3     0     0     0     0     0     0     4     0     0     0
#> 4     0     0     0     0     0     0     0     0     3     0
#> 5     0     0     0     0     0     0     0     0     0     2

but I feel like there should be an easier or "tidier" way to make all of the assignments at once, especially if there are many more than 5 items. Suppose that we know we won't get the same cell overwritten or anything (index combinations won't be duplicated), so the cell being modified will not change depending on what cycle you are on. You could call this question "vectorising assignment".

Rui Barradas · Accepted Answer · 2018-08-22 07:34:13Z

6

This can be done with no loops at all, be them for or *apply loops.
The trick is to use an index matrix. But since this only works for target objects of class matrix, coerce the tibble or data.frame to matrix and then coerce back.

I will repeat the inclusion of the data creation code, with @Ronak's solution, to make the code self contained.

inx <- cbind(rows, match(cols, names(df1)))
df1 <- as.matrix(df1)
df1[inx] <- vals
df1 <- as.tibble(df1)

identical(df, df1)
#[1] TRUE

Data creation code.

set.seed(1234)
cols <- sample(letters[1:10], 5)
vals <- sample.int(5)
rows <- sample.int(5)
df <- matrix(rep(0L, times = 50), ncol = 10) %>%
  `colnames<-`(letters[1:10]) %>%
  as_tibble

df1 <- df
mapply(function(x, y, z) df[x, y] <<- z, rows, cols, vals)

answered Aug 22, 2018 at 7:34

Rui Barradas

78k8 gold badges41 silver badges75 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Jaap Over a year ago

no need to convert df1 to a matrix imo; df1[inx] <- vals works on a dataframe as well

Rui Barradas Over a year ago

@Jaap You are right, I had tested it with the tibble. In the data creation, change as_tibble to as.data.frame and the matrix index works.

Ronak Shah · Accepted Answer · 2018-08-22 07:08:40Z

6

One way using mapply would be:

mapply(function(x, y, z) df[x, y] <<- z, rows, cols, vals)
df

#      a     b     c     d     e     f     g     h     i     j
#  <int> <int> <int> <int> <int> <int> <int> <int> <int> <int>
#1     0     0     0     0     5     0     0     0     0     0
#2     0     0     0     0     0     0     0     2     0     0
#3     0     0     0     0     0     1     0     0     0     0
#4     0     4     0     0     0     0     0     0     0     0
#5     0     0     0     0     0     0     0     0     3     0

You can read more about <<- operator here.

data

set.seed(1234)
cols <- sample(letters[1:10], 5)
vals <- sample.int(5)
rows <- sample.int(5)

answered Aug 22, 2018 at 7:08

Ronak Shah

391k20 gold badges173 silver badges237 bronze badges

Collectives™ on Stack Overflow

How can I avoid a loop when replacing values in a dataframe by row-column indexing?

2 Answers 2

2 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related