Grouping adjacent cells by given number

[MarcoMarchetti]

Hello everybody,

we have a vector polygon made by 1'206'198 (500x500m) cells.

We need to make groups of 9 adjacent cells (3x3). The selected cells must belong to one and only one group.

A complication is that our grid is not regular but it is a bit distorted in order to account for the sferical and irregular terrestial surface.

Is ther any tool or proven procedure to achieve such a result?

Thanks in advance!

[JakeSkinner]

Hi Marco,

Not sure if this will give you the results you are looking for, but take a look at the Block Statistics tool.

[GerryGabrisch]

What do you mean by 'not regular'?  Are the cells not all perfect squares, is the distribution of the cells not rectangular, or are there cases where the 3 x 3 groupings will leave left-overs? 

If the  cells are squares and the distribution of squares lends itself to 3 x 3 groupings across the surface then...
Use the create fishnet tool to generate a feature class of cells that are 1500m x 1500m.  Then use a spatial join to populate the original cells with the OID of the 1500 x 1500 cells.

[MarcoMarchetti]

What do you mean by 'not regular'?  Are the cells not all perfect squares, is the distribution of the cells not rectangular, or are there cases where the 3 x 3 groupings will leave left-overs? 

If the  cells are squares and the distribution of squares lends itself to 3 x 3 groupings across the surface then...
Use the create fishnet tool to generate a feature class of cells that are 1500m x 1500m.  Then use a spatial join to populate the original cells with the OID of the 1500 x 1500 cells.

Exactly, the cells are not perfect squares and their distribution is not rectangular, it is the Italy shape, so left-overs are likely to be created.

Do you have any idea in this case?