Near (Error: 999999) Failed to execute (near)

[GrantWest]

I have a shapefile of roughly 1.3 million points, and I am trying to calculate the distance to the closest of 860 polygons (also shapefile) for each, using the Near tool. I ran a Near calculation and continue to get this error message that it failed to execute the function. I tried doing it again using the "Repair Geometry" tool before the "Near" tool. It repaired my geometry in about 15 minutes. After repairing, it took much longer for the operation to kick back an error message, but eventually it gave me the same one. What are the probable causes of this issue and how can I work around it? Do I need to split up my data? I'm running a 2.8 GHz processor and 8 GB RAM.

A secondary question:
My point data set are geocoded addresses of students given in decimal degrees format. The data came to me from Dun & Bradstreet, though I had nothing to do with the data acquisition. How can I know what geographic reference system these decimal degree coordinates come in for when I create a my shapefile from XY data? At this point, I have them as simply NAD83. Does anybody know how I can "know" the reference system or what kind of assumptions I should make?

-Grant

[markdenil]

In my experiance, any operation on that many points is likley to run into a wall.
try splitting off a chunk and running Near on that.

You can create a set of Feature Layers based on the $recno, excuse me, the ObjectID and run each in turn.

The datum issue I think, will hinge on how close together the polygons lie.
The difference in lat/longs will likely be, at most, in hundreds of meters.
If your polygons are tens of kilometers apart or more, I doubt a datum mismatch will be significant.

[GrantWest]

Thank you.  I just exported a selection of 100,000 points to a new shapefile and ran it again.  This time it worked.  I am confident you are right that differences due to the datum will not be significant.  I just wanted to dot my I's and cross my T's.  My polygon layer is projected in UTM while I set my points to be geographic with NAD83.  I ended up projecting my points prior to using the Near tool to get my output in meters. 
 
Currently, I am working with a dummy data set on my workstation because the real point data I will be using for my analysis are government-protected, sensitive location data.  I am developing a turnkey process to allow a government agency to run this on the real data on a mainframe computer in our state capital.  Will it likely still be necessary to split the data into chunks to run the operation on the real data?