Thank you for help!
It's actually too complicated for picking tile by ID for a non programmer like me.
I guess one day there will be simple condition like :
add new condition > on collision with another object > tileMap (indicate tile ID in text field below)... and that's it!
Until this kind of improvement, I will use tile map only for decoration.
The tilemap is not really that hard to use once you understand the basics of how it is constructed.
Image (1): Just using the basic tilemap in C2. You can look at a tilemap as being just a collection of images combined into 1 image. The reason to use a tilemap is because you can reuse a lot of graphic and reduce the performance hit from using a lot of sprites. So instead of adding lots of objects you simply add a tilemap and reuse the graphic over and over. There are some other benefits as well, but to keep it simple ill just focus on the basics.
Image (2): Shows a single tile in the tilemap, which current size is set to 32 x 32 px, this is equal to adding a sprite and setting its width and height to 32px. But since the tilemap split the image into tiles, you need to tell C2 what size these tiles are, so it knows how many tiles the tilemap should contain.
Image (3): Once the size of each tile have been set, C2 split the image into tiles and give each of them an ID, so its easy for you to see which graphic belongs to what tile. And it makes it possible for C2 to work with the tiles. This ID identify what image is related to a given tile. And can be compared to you making 2 sprites and calling one of them "Image_1" and the other "Image_2".
Image (4): Guess this might be what confuse you a bit. But the tilemap object can be compared to adding a lot of sprites next to each other. But instead of you having to add all of them, you just add the tilemap object and C2 will then split it up for you in the tile sizes that you set in Image (2). The tilemap object, that being the one in the C2 viewport will be split into areas that match the size of the tile width and height.
So a tilemap that is 320x320 px and have a tile width and height of 32 x 32 px will have 10 tiles in X and 10 tiles in Y. And each of these tiles will have an index so C2 know which tile is which. Looking at Image (4) you can see the tile coordinates, which always starts at 0,0 in the top left tile. This coordinate have nothing to do with the X,Y value in the normal coordinate system that C2 uses. So you can place the tilemap at X: 500 and Y: 500 in the viewport, but the top left tile in the tilemap will always be (0,0).
Image (5): Shows a tile placed on the tilemap, this tile have the tile position of (2,1) and looking at the overview in image (3) you can see that the ID of the tile is 0.
If this is makes sense to you, you can see that you can't work with a tile ID alone. Because it simply relates to what type of tile it is. You have to first tell C2 which tile coordinate you want to test.
And you do that by using the condition "Compare tile at":
To test it:
So you just have to remember that Tile ID and Tile coordinate cant work without each other. The Tile ID is the tile graphic in the tilemap, and the tile coordinate is on the tilemap it self.