Quickest Way to Spatial Join 250000 Features

AndrewJoseph

Hello,

I am trying to look at neighborhood landuse charactersitics within a quarter mile around all landuse polygons in a database of roughly 250,000 polygons.

I'm using an Intel CORE I7 Vpro with 64 bit and a total of 8 processors with arc10, but even it seems unable to manage spatial joining this dataset even after I have split the file up into feature classes of only 35,000 each! Three days of my existence have been wasted on this problem. It seems to get through the entire process but then cannot write the attribute table.

So, I was wondering if anyone had any ideas about the best way to spatial join a dataset of this size, and what factors besides size of dataset affect spatial join speed? 

On a related note, is it faster to first create quarter mile buffers around each landuse polygon, or simply specify a quarter mile  search radius in the join dialog? I also have a dataset consiting of points that are 660 ft apart along the street network, would it be quicker just to join landuses a quarter mile from these points than the landuse polygons themselves?

ColinLang1

This is the tiny piece of information I've been searching for, for months.  Screw the spatial join, Intersect produces a similar enough result, with a massive performance improvement.  I just made a script that takes less than 2 minutes and replaces my previous effort that used spatial joins, that took 16 hours.

Thanks!