Finding a random spot that is free...

Ceril

Greetings.

Well. The outset is this:

TblSquaresLayout //Contains all squares for the map.
-ID
-MapID
-OwnerID

TblMapInfo //Name of the map, its layout size in x * y
-ID
-Name
-SizeX
-SizeY
-Wrap

Now on to the question:
How to have mysql return one Random TblSquaresLayout.ID that have a OwnerID=0 and no other squares with OwnerID>0 in a 2 square radius.
Lets say MapID = 1 and that MapInfo for #1 is: 1, Test, 50,30,1 -> Map number one is named test and its layout is 50 wide and 30 high and its wraparound, meaning that its... squares are actually positioned upon a sphere surface.

Its the Random and the Wraparound part that have made me dump highly structured pop this day.

Roger_Ramjet

Well the first part of your problem is easy, you use ORDER BY RAND() LIMIT 1 to return a random records from the selected results.

Roger_Ramjet

Selecting those results appears to be the main problem. Wrap-around hmm!

Hang on a minute. How are the squares indentified then? I see no x and y co-ordinates!! How do you identify where each square is on the map and therefore which other squares are next to it?

What have you left out in your post 😕 Real table layouts if you want people to help you :mad:

Ceril

Hmm... that might be the trick. Should be alot easier if I actually Give the squares their X,Y coords directly instead of having to calculate them from SizeX which probably was a lousy idea.
Doh... This map thingy aint my stuff 😃 And it is the actual table structure, its nothing fancy atm, Oh I forgot to type out color.

Ok, reworked the TblSquaresLayout to contain
-ID
-MapID //What MapInfo.ID does this map belong to
-Color //Varchar 7, hex color (#000000)
-X
-Y
-OwnerID

It got alot easier to grab a piece of the map centered around a certain square instead of the and and and and and query before
But now remains how to do a wraparound which I'am clueless to. Iam waiting with trying the RAND() until I got wraparound down, or is faced with defeat.

Thanks.
/C

Edit:
Ok... confusion arises. First I was dumb and made a new 100100 map. Then I made one that is 66 with OwnerID=1 for coords 6,1 and 1,6 meaning only 3,3 and 4,4 should be valid coords which I cant get them to be -_-

Edit X: I think I might just be to tired to actually think straight about 😃

Ceril

Today after some sleep:
Basicly what I'am looking for should be something like this:
SELECT ID, X,Y FROM TblSquaresLayout as S WHERE S.OwnerID=0 AND NOT EXISTS
(SELECT * FROM TblSquaresLayout WHERE X>=S.X-2 AND X<=S.X+2 AND Y>=S.Y-2 AND Y<=S.Y+2 AND OwnerID>0)
In my mind I ask this question: Give me Id, X and Y From SquaresLayout where there is no owner and no squares within the coords given have an owner.

But as you all might guess, it does not work 😃 So I guess I'am out in the clear blue sky no gold veins to follow.

Roger_Ramjet

OK, so wrap-around: what do you mean by that?

If you are thinking of the usual map of the world and the way that that wraps around a globe, then that is a Mercator Projection map and all you need is that the squares wrap on the Y-axis.

If you want all the edges of the square map to wrap into each other then that is not a globe, that is a Pyramid. You can of course distort the base and walls to make it globular in shape, but mathematically it is still a pyramid.

Both of the above result in 'squares' being distorted in shape when wrapped around a glode, but that does not effect the game I'm sure.

If you want a real 3D globe then the 'squares' have to be haxagons and you have 6 sides and 3 axes, X Y Z. Then you can really talk about wrapping areas around a globe.

So which is it?

Ceril

All this for something that will be used once or twice in the end 🙂
Darn, didnt think about that little fact when it comes to wrapping something up. The easiest way and not overly complex is to got for Y axis wrap as to not make things more confusing than they need be, however not I got curious as to just how simulate a 3d globe effect 2d, if you got any nice pointer to the math behind it include it please.

Thanks 🙂

Roger_Ramjet

Well, to see one way of doing it, make a numbered grid on a piece of paper, turn it over and fold each corner into the centre. You will see that you now have an unbrocken surface that with a little imagination, and stretching, could be inflated into a globe. Each edge wraps to itself thus

6x6 grid
along first edge, 1,1 wraps to 1,6 : 1,2 to 1,5 : 1,3 to 1,4
along second edge 1,1 to 6,1 : 2,1 to 5,1 : 3,1 to 4,1
along 3rd edge 6,1 to 6,6 : 6,2 to 6,5 : 6,3 to 6,4
along 4th edge 6,6 to 1,6 : 5,6 to 2,6 : 4,6 to 3,6

Probably a maths formula in there somewhere, and bound to have been worked out by someone so try a google for it.

If you stick with the standard mercator projection then each 'square' along the top and bottom rows should really wrap to any other square on that row.

You could go for a cube and have a uniform shape and size of 'square'. The pyramid solution is just half a cube anyway. Still need 3 axes then to cover the face of the cube, and the wrap-mapping would have to be a static table rather than a calculated function.

For a realistic globe then hexagons and 3 axes will give the best fit. I don't know what the formula is for the number on a sphere, bound to be a limited number of possible totals for a complete sphere, so google for it.

Roger_Ramjet

Of course it is possible to wrap on both axes, but what you get is a doughnut ring. Now there are those who maintain that this is the true shape of space-time anyway so there is no reason your virtual universe should not be that shape as well. Especially as you are not planning to display the 'globe' graphically except as a flat map.

The formula for the wrapping is very straight forward: in pseudo-code

X and Y = size of map
x = x of square
y = y of square

if (x-2) < 1 then X+(x-2) else (x-2)

if (x+2) > X then (x+2)-X else (x+2)

likewise for y