Use tapply(dataframe , index, function) in R giving as argument to the function 2 columns

Question

I would like to use the tapply() function on a dataframe, grouping the rows with the indexing. My problem is that the argument I would pass to the function is not a single column, but a pair of columns. This beacause the 2 columns of the data frame represent x-y points, which are intended as couples. Running tapply(dataframe , indexes , function) gives me the error that indexes has length different from tapply. How can I solve this? Thank you!

akrun akrun · Accepted Answer · 2021-04-17T21:55:12

If there are more than one column to be summarised, use aggregate instead of tapply (as tapply works for a single column)

aggregate(.~ indexes, transform(df1, indexes = indexes), FUN = yourfun)

Or another option is by

by(df1, list(indexes), FUN = yourfun)

Or it may be more flexible with tidyverse

library(dplyr)
df1 %>%
    group_by(indexes) %>%
    summarise(across(c(x, y), yourfun), .groups = 'drop')

Using a small reproducible example

indexes = rep(1:2, c(3, 2))
by(mtcars[1:5, 1:5], indexes, FUN = sum)

Use tapply(dataframe , index, function) in R giving as argument to the function 2 columns

1 Answers