Want to share your content on R-bloggers? click here if you have a blog, or here if you don’t.
Happy December, R friends!
One of my favorite traditions in the R community is the Advent of Code, a series of puzzles released at midnight EST from December 1st through 25th, to be solved through programming in the language of your choosing. I usually do a few of them each year, and once tried to do every single one at the moment it released!
This year, I know I won’t be able to do it daily, but I’m going to do as many as I can using just data.table solutions.
I’ll allow myself to use other packages when there isn’t any data.table equivalent, but my solutions must be as data.table-y as possible.
I’m going to abuse the blog post structure and update this file throughout the week.
December 1st
Part One
d1 <- fread("day1_dat1.txt")
d1[, V1 := sort(V1)] d1[, V2 := sort(V2)] d1[, diff := abs(V1-V2)] sum(d1$diff)
Part Two
d1[, similarity := sum(V1 == d1$V2)*V1, by = V1] sum(d1$similarity)
December 2nd
Part One
d1 <- fread("day2_dat1.txt", fill = TRUE)
check_report <- function(vec) {
vec <- na.omit(vec)
has_neg <- vec < 0
has_pos <- vec > 0
inc_dec <- sum(has_neg) == length(vec) | sum(has_pos) == length(vec)
too_big <- max(abs(vec)) > 3
return(inc_dec & !too_big)
}
d1t <- transpose(d1) deltas <- d1t[-nrow(d1t)] - d1t[2:nrow(d1t)] res <- apply(deltas, 2, "check_report") sum(res)
Part Two
test_reports <- function(dat) {
deltas <- dat[-nrow(dat)] - dat[2:nrow(dat)]
res <- apply(deltas, 2, "check_report")
res
}
res <- test_reports(d1t)
for (i in 1:nrow(d1t)) {
res <- res | test_reports(d1t[-i,])
}
sum(res)
Just for fun
I found the use of apply deeply unsatisfying, even though it was fast, so just for fun:
d1t <- transpose(d1) deltas <- d1t[-nrow(d1t)] - d1t[2:nrow(d1t)] is_not_pos <- deltas <= 0 is_not_neg <- deltas >= 0 is_big <- abs(deltas) > 3 res_inc <- colSums(is_not_neg | is_big, na.rm = TRUE) res_dec <- colSums(is_not_pos | is_big, na.rm = TRUE) sum(res_inc == 0) + sum(res_dec == 0)
Yay. 🙂
Related
