4 ms·

Things I learnt about Julia during Advent of Code

9 points by mewfree 5y ago

snicker7 5y ago
Julia matrices are laid out as column-major (like Fortan, unlike C). So, it is better to get into the habit of thinking of a matrix more as a vector of columns than a vector of rows.
dunefox 5y ago
I also did AoC in Julia and it was pretty great. I did a few days in Kotlin as well and the value of good tooling is hard to overstate. This is one of my biggest problems with Julia, otherwise it's a very good language.
stellartux 5y ago
There are few easier ways built in to do some of the file functions you've made, instead of `open(f -> read(f, String), "input")` you can use `readchomp("input")`. There's also `readlines` which is equivalent to `open(f -> read(f, String), "input") |> f -> split(f, "\n")`, and `eachline` which is like `readlines` but returns a generator instead of an array. There are curried forms of the operators that might make it easier to write functional code, like instead of `x -> x == "0"` you can write `==("0")`. Another useful thing is you can broadcast pipes, so instead of `x |> xs -> f.(xs)` or `x |> Map(f)`, you can write `x .|> f`. It's funny that the first section is about everything loading as a matrix and not knowing the easy way to load things as a vector, I had the opposite problem and every problem that has a matrix I've ended up making a giant vector then reparsing the first line to figure out what dimensions to reshape it to. I'll have to try `readdlm`, I'm currently working through the older AoC years in Julia.
- dunefox 5y ago
  For better piping there's https://c42f.github.io/Underscores.jl/stable/ https://c42f.github.io/Underscores.jl/stable/. It eliminates having to explicitly write anonymous functions.
kwertzzz 5y ago
Nice article! One can you also use reduce(hcat,n) instead of hcat(n...). The function reduce(hcat,n) should be faster as there is optimized implementation of it: https://github.com/JuliaLang/julia/blob/89a51fb98485219680c7106e94cf75fd8069f3f8/base/abstractarray.jl#L1375-L1379 https://github.com/JuliaLang/julia/blob/89a51fb98485219680c7... This feature was discussed here: https://github.com/JuliaLang/julia/issues/21672 https://github.com/JuliaLang/julia/issues/21672

9 points by mewfree 5y ago

Julia matrices are laid out as column-major (like Fortan, unlike C). So, it is better to get into the habit of thinking of a matrix more as a vector of columns than a vector of rows.

dunefox 5y ago

I also did AoC in Julia and it was pretty great. I did a few days in Kotlin as well and the value of good tooling is hard to overstate. This is one of my biggest problems with Julia, otherwise it's a very good language.

stellartux 5y ago

There are few easier ways built in to do some of the file functions you've made, instead of `open(f -> read(f, String), "input")` you can use `readchomp("input")`. There's also `readlines` which is equivalent to `open(f -> read(f, String), "input") |> f -> split(f, "\n")`, and `eachline` which is like `readlines` but returns a generator instead of an array. There are curried forms of the operators that might make it easier to write functional code, like instead of `x -> x == "0"` you can write `==("0")`. Another useful thing is you can broadcast pipes, so instead of `x |> xs -> f.(xs)` or `x |> Map(f)`, you can write `x .|> f`. It's funny that the first section is about everything loading as a matrix and not knowing the easy way to load things as a vector, I had the opposite problem and every problem that has a matrix I've ended up making a giant vector then reparsing the first line to figure out what dimensions to reshape it to. I'll have to try `readdlm`, I'm currently working through the older AoC years in Julia.

dunefox 5y ago
For better piping there's https://c42f.github.io/Underscores.jl/stable/ https://c42f.github.io/Underscores.jl/stable/. It eliminates having to explicitly write anonymous functions.

kwertzzz 5y ago

Nice article! One can you also use reduce(hcat,n) instead of hcat(n...). The function reduce(hcat,n) should be faster as there is optimized implementation of it: https://github.com/JuliaLang/julia/blob/89a51fb98485219680c7106e94cf75fd8069f3f8/base/abstractarray.jl#L1375-L1379 https://github.com/JuliaLang/julia/blob/89a51fb98485219680c7... This feature was discussed here: https://github.com/JuliaLang/julia/issues/21672 https://github.com/JuliaLang/julia/issues/21672