Awk, some notes

Inspired in a similar post. I though I would share my notes on Awk that I took while I was doing adventofcode last year.

Here “Bad” are things that will cause bugs on your code. And “Ugly” are annoying things that once your accept them might make you think different approaches to your problem.

The Good

  • a regex pattern matching for each line can carry you longer that I though possible

  • shines when parsing text files

  • if you have been playing around with pattern matching and guards on a functional programming language, you might feel oddly at home. Either with line matching or by using switch/case which also supports regexes.

  • Ternary operators might also become your best friend (which might make you reflect on the shape they take in other programming languages)

The Bad

  • typos on variable names are painful due awk’s behavior of defaulting uninitialized variables to 0 or “”

  • add it to the fact that if you even touch (reference) a element that doesn’t exists on an array it will be created

  • per above point, you might want to create an “getter” for arrays

    function map_at(    x,y) {
      return (x in map && y in map[x]) ? map[x][y] : "?"
    }
    
  • at the end of the day it is a weakly typed language, so even if it looks like number and you can operate on it like a number, if you compare it with something else it might act like a string!!!

The Ugly

  • lack of a extensive stdlib is also a pain, especially with arrays and strings

  • conditions on for loops get evaluated each iteration (this is not awk’s fault, I am just too functionally immutable poisoned)

  • lack of another data type besides associative arrays makes is so you have to get creative and create wacky indexes to imitate complex dictionaries, like:

    mapping["seeds"][2]["distance"] = 42
    
2 Likes