Converting multiple polygons to equal sized rasters

[JohnBencina1]

I am working with the block-group level Tiger geodatabase from the US Census Bureau. What I wanted to do was convert the polygons into 10x10mi raster cells. I have multiple rasters for several different Census variables. My question is if there is a more efficient process to what I am doing today.

I need the cells to be uniform because the Census rasters are being combined with other rasters. So first I created a Fishnet Grid over the US. I then perform a Spatial Join with the fishnet layer & the census layer taking the average of the values because one cell could contain more than one block-group polygon. Once joined, I run Polygon to Raster with the same cell-size to output a series of rasters all with identically shaped & positioned raster cells.

The Spatial Join process takes the longest so I'm wondering if someone has a more efficient approach to this problem.

[DanPatterson_Retired · Accepted solution]

Your files are large I suspect, maybe your data are on a network and the computer just meets the minimum requirement to run Map or Pro.  Anything I have missed? If it works, that seems like a reasonable workflow but more importantly, how much time are you hoping to shave off what you are doing now?

[SteveLynch]

How many columns and rows in the output raster?

How many rows (or features) in the input polygon feature class?

What raster format are you using: fgdb, grid, img, tiff?

Polygon feature class in what format: .shp or fgdb?

What version of ArcGIS?

[JohnBencina1]

Yeah I suspect it's just large data files. It just seemed subjectively "sluggish". I wanted to make sure I wasn't missing something obvious since there's always more than one way to approach a problem like this.

[DanPatterson_Retired]

consider it 'reflection and coffee' time rather than sluggish

[MicahBabinski]

Hey John,

Dan and Steve raise good points. To expand on those points:

1. Work in file geodatabase rather than shapefile (shapefiles are difficult to index efficiently which will slow down your process)
2. Use The in_memory workspace if at all possible - it'll run faster because the data need not be written to your hard drive or network location
3. Process the data locally on your C drive if in_memory doesn't work. You can pick your own place or use the Scratch GDB, which will do the work in your user profile.

Finally, one question: do you need the spatial join? Depending on your cell size, the characteristics you are trying to portray, and your accuracy/precision constraints, you may just be able to run polygon to raster directly on your census blocks and skip the whole spatial join step.

Good luck! Performance issues can be tough to handle but they make for good learning experiences.

Micah

[AbdullahAnter]

What about convert  Polygon to Raster—Help | ArcGIS Desktop ,then use Split Raster—Help | ArcGIS Desktop ?