CODEWITHSUNDEEP
rlistdataframepurrr
MayaGans
MayaGans

Reputation: 1845

dynamically set unique column names in list of dataframes [R]

I'm trying to generalize the code here for more than one repeating column: rename list of dataframe columns to mimic joined suffixes

I have a dataframe with columns name that are the same amoung different data frames in a list I want to use the same patters as reduce(left_join, suffix = c("_x", "_y"), by="inAll") to create new names

PATTERN:

library(dplyr)
library(purrr)
library(stringr)

dd <- list()
dd$data <- list(
  ONE = data.frame(inAll = c(1.1,1.2,1.3), inAll_2 = c(1.4,1.5,1.6), inSome = c(1.7,1.8,1.9), only_one = c(1.10, 1.11, 1.12)),
  TWO = data.frame(inAll = c(2.1,2.2,2.3), inAll_2 = c(2.4,2.5,2.6), inOthers = c(2.7,2.8,2.9)),
  THREE = data.frame(inAll = c(3.1,3.2,3.3), inAll_2 = c(3.4,3.5,3.6)),
  FOUR = data.frame(inAll = c(4.1,4.2,4.3), inAll_2 = c(4.4,4.5,4.6), inOthers = c(4.10, 4.11, 4.12), inSome = c(4.7,4.8,4.9)),
  FIVE = data.frame(inAll = c(5.1,5.2,5.3), inAll_2 = c(5.4,5.5,5.6)),
  SIX = data.frame(inAll = c(6.1,6.2,6.3), inAll_2 = c(6.4,6.5,6.6), inOthers = c(6.7,6.8,6.8))
)

Desired output

dd$data2 <- list(
  ONE = data.frame(inAll = c(1.1,1.2,1.3), inAll_2_x = c(1.4,1.5,1.6), inSome_x = c(1.7,1.8,1.9), only_one = c(1.10, 1.11, 1.12)),
  TWO = data.frame(inAll = c(2.1,2.2,2.3), inAll_2_y = c(2.4,2.5,2.6), inOthers_x = c(2.7,2.8,2.9)),
  THREE = data.frame(inAll = c(3.1,3.2,3.3), inAll_2_x_x = c(3.4,3.5,3.6)),
  FOUR = data.frame(inAll = c(4.1,4.2,4.3), inAll_2_y_y = c(4.4,4.5,4.6), inOthers_y = c(4.10, 4.11, 4.12), inSome_y = c(4.7,4.8,4.9)),
  FIVE = data.frame(inAll = c(5.1,5.2,5.3), inAll_2_x_x_x = c(5.4,5.5,5.6)),
  SIX = data.frame(inAll = c(6.1,6.2,6.3), inAll_2_y_y_y = c(6.4,6.5,6.6), inOthers = c(6.7,6.8,6.8))
)

How to get there:

Open to entirely different ideas!!!

new_names <- function(data, toExclude) {
  # calculate the number of suffixes we need per column
  nnames <- map(data, ~colnames(.x)) %>% 
    unlist() %>% 
    tibble() %>% 
    rename("names" = 1) %>% 
    group_by(names) %>% 
    count() %>%
    filter(!names %in% toExclude)

  suffixes <- map(nnames$n, ~strrep(rep(c('_x', '_y'), .x/2), rep(seq_len(.x/2), each = 2)))
  map2(nnames$names, suffixes, ~paste0(.x, .y))

  # .....somehow apply these to the right columns?
}

Function specifying data frame and column to exclude which should yield the desired output:

new_names(dd$data, "inAll")

Any help, even with just what logic to use to get my desired end result would be really appreciated, thank you!

Upvotes: 2

Views: 296

Answers (1)

akrun
akrun

Reputation: 886948

Here is one option, where we extract column names from the list, change the names based on the how it is duplicated after splitting by common names, then relist it back to list again and change the column names of the original list using map2 with the changed names

library(purrr)
library(dplyr)
library(stringr)
inp <- dd$data
lst1 <- map(inp, names)
nm1 <- unlist(lst1)
i1 <- nm1 != 'inAll'
lst2 <- split(nm1[i1], nm1[i1])
i2 <- lengths(lst2) > 1
lst2[i2] <- map(lst2[i2], ~  { 
                    i3 <- (seq_along(.x)-1) %/% 2 + 1
                     i4 <- i3 %in% names(which(table(i3) == 2))
                     n <- ceiling(length(.x[i4])/2)
                     .x[i4] <- str_c(.x[i4], strrep(rep(c('_x', '_y'), n),
                                  rep(seq_len(n), each = 2)))
                     .x  
                      })
nm1[i1] <- unsplit(lst2, nm1[i1])
out2 <- map2(inp, relist(nm1, skeleton = lst1), set_names)

-checking with OP's output

out <- dd$data2
identical(out, out2)
#[1] TRUE

Upvotes: 2

Related Questions