I have a grid where in each individual coordinate location it holds information on:
The ground type
The wall type
IF - There is a wall
The object type
IF - There is an object
AND IF - There is not a wall
The pipe type
IF - There is a pipe
I am not sure what your question is here… just about data structures in general or did you have a specific issue?
How to report problems productively in the Unity3D forums:
Help us to help you.
I wanna know the best way of storing grid data to be accessed by many scripts, the data would be in a matrix like
[ [X, X, X]
[X, X, X]
[X, X, X] ]
Where X is a struct with all the information for that spot
Do you know how many data you will have at the end ?
You can store it for example in a file where each line is one object with the comma separated values.
So you can read that file line by line, create a new object for each line, split the values by , (comma) and set the the values to the object.
But it´s not the best way if you have thousands of lines/data/objects.
Hope i understood it correct and the answer helps you.
The best way will always be dependent on details such as how long you want to store it and how massive it is, how sparse it is, etc.
But this just seems like a very straightforward “cells in a grid with stuff inside 'em.”
Generally store things in a class:
https://docs.microsoft.com/en-us/dotnet/csharp/programming-guide/classes-and-structs/classes
And you can either use a Dictionary indexed by a 2-dimensional custom key:
https://docs.microsoft.com/en-us/dotnet/api/system.collections.generic.dictionary-2?view=netcore-3.1
https://stackoverflow.com/questions/6999191/use-custom-object-as-dictionary-key
Or just store it in a 2D array:
https://docs.microsoft.com/en-us/dotnet/csharp/programming-guide/arrays/multidimensional-arrays
or if you prefer, a jagged 2D array:
https://docs.microsoft.com/en-us/dotnet/csharp/programming-guide/arrays/jagged-arrays
Regardless of any choice above, FAR more important is to immediately push past the first steps of this, choose an approach (any approach is better than analysis paralysis) and start working with actual data so you can begin to evaluate which way might work best in your exact context, which none of us have access to here.
1 Like
The dictionary choice seems like the best option, how would I create a 2-dimensional custom key?
Did you read the stackoverflow link? I’m not sure I could explain it any better. What part are you having difficulty with?
By assuming you know what you’re doing, here are your options:
a) you compute a ‘flat’ index
b) you construct a value tuple
c) you compute a hash with a custom hash function
d) you compute a Cantor pairing value
Without much hesitation, I’d always pick (a) in your case
This is how to implement these
a)
int index = y * width + x; given that your width is fixed for each row, you end up with unique indice
to reverse a flat index you have to do the following
y = index / width; (<- this is an integer division, not an ordinary one!!)
x = index % width; (<- this is a modulo remainder division)
b)
(int x, int y) key = (x, y);
C# will generate a proper hash code for each value combination, but I’m not sure if it guarantees uniqueness for a very large set
you don’t have to reverse anything, the source values are accessible at all times, by doing
key.x or key.y
c)
the most banal hash function would be
int hash = (y << 16) | x;
you will run into hash collisions if your x or y can be equal to or greater than 2^16 for this particular scenario
(look up into proper hash generators online)
you can’t normally reverse a hash key cheaply, but in this banal scenario you actually can by doing
int x = hash & 0xf, y = hash >> 16;
d)
Cantor pairing
int pairing = unchecked(((x + y) * (x + y + 1) >> 1) + y);
uniqueness is mathematically guaranteed (unless you hit an overflow)
for reversing check the link