My very tidy solution to day7 in the advent of code

Advent_of_code_2024_day07_part2.R

library(gtools)
library(stringr)
library(dplyr)
library(tidyr)
d7 <- readLines("day07.txt")
#assumes you kept the resut of part1 to save work
d7p1 <- as.numeric(readLines("d7p1result.txt"))
p2data <- d7[d7p1 == 0]
p2sums <- as.numeric(gsub(":.*","",p2data))
p2input <- gsub(".*: ","",p2data)
p2results <- data.frame(dataID = p2input, expected_results = p2sums)
p2operationscount <- str_count(p2input, " ")
# 1 is addition, 2 is multiplication, 3 is concatination, 0 is hold
permus <- function(x, x_index=p2input,y_ops=p2operationscount, paddingvalue=max(p2operationscount)){
  x_index <- x_index[x]
  y_ops <- y_ops[x]
  plustimes = permutations(n=3,r=(y_ops), v=1:3, repeats.allowed=TRUE)
  #we checked perms without concat in part1
  has_concat = which(apply(plustimes, 1, function(x){sum(x == 3)})>0)
  wanted_rows = plustimes[has_concat,1:ncol(plustimes)]
  padding <- matrix(data=rep(0,nrow(wanted_rows)*(paddingvalue-ncol(wanted_rows)+1)), 
                    ncol = (paddingvalue-ncol(wanted_rows)+1))
  padded <- as.data.frame(cbind(wanted_rows,padding))
  padded$dataID = x_index
  return(padded)
}
p2_data_ops_L <- lapply(1:length(p2input),permus)
p2_data_DF <- bind_rows(p2_data_ops_L)
rm(p2_data_ops_L)
p2_operations <- p2_data_DF |> select(dataID,V1:V11)
p2_numbers <- p2_data_DF |> select(dataID) |> 
  separate_wider_delim(dataID, delim=" ", names=paste0("N",1:12),
                       too_few = "align_start", cols_remove = FALSE)
p2_numbers <- p2_numbers %>%
  mutate(across(starts_with("N"), as.numeric))
options(scipen = 999)
rolling_outcome <- p2_numbers[[1]]
for(i in 1:11){
  operations_i <- p2_operations[[i+1]]
  new_numbers_i <- p2_numbers[[i+1]]
  rolling_outcome[operations_i == 1] <- rolling_outcome[operations_i == 1] +
    new_numbers_i[operations_i == 1]
  rolling_outcome[operations_i == 2] <- rolling_outcome[operations_i == 2] *
    new_numbers_i[operations_i == 2]
  rolling_outcome[operations_i == 3] <- as.numeric(paste0(rolling_outcome[operations_i == 3],
                                               new_numbers_i[operations_i == 3]))
}
outcometest <- data.frame(testresult = rolling_outcome, dataID=p2_operations$dataID)
foundresults <- outcometest |> left_join(p2results, join_by(dataID)) |> 
  filter(testresult == expected_results) |> 
  distinct()
sum(d7p1) + sum(foundresults$expected_results)