NSDI Data Update has Different Spatial Index Grid - Will This be an Issue?

[LauraSilsbee1]

I am using the ArcGIS 10.0 Feature Compare tool to review schema changes between our exisiting SDE NSDI dataset and a new update we just received.  I am seeing this error on every feature class:  "GeometryDef grid size 1 does not match"

If we replace our existing dataset with the new one what issues might we see because of the different spatial index?

Thanks,
Laura

[BarryConner]

I am using the ArcGIS 10.0 Feature Compare tool to review schema changes between our exisiting SDE NSDI dataset and a new update we just received.  I am seeing this error on every feature class:  "GeometryDef grid size 1 does not match"

If we replace our existing dataset with the new one what issues might we see because of the different spatial index?

Thanks,
Laura

I'm very interested to see an authoritative answer to your question, Laura.  I've been seeing the same thing using Table Compare in ArcGIS 10.0. 

I found your posting when trying to resolve my own concerns about this issue.

Remarkably, I just got this same result upon comparing an existing (base) table with a newly created 'clone' of that table -- in the same database and in the same feature dataset; relying on that 'parent' spatial environment to serve as projection and datum; simply importing the original schema and loading the data from the already-existing table. There should have been only one difference noted: I had increased the string length of one text field from 50 characters to 255 characters in the newly-imported schema before finalizing the feature class creation and actually loading the data. It seems there should have been no change whatsoever, to the new table, spatially.

Report of Table Compare is attached.

Barry Conner
GIS Analyst
City of Huntsville, Texas

[BarryConner · Accepted solution]

Laura, I hope this might help you:  I've gotten a solution to the 'GeometryDef grid size 1 does not match' error from my tabular attribute comparison of the two identical point feature classes.

At the suggestion of my GIS Manager, I went to the Indexes tab on the feature class properties in ArcCatalog for both of these features.  Under 'Spatial Index', I performed an automatic 'Recalculate'.  The default, automatic recalculation of the spatial index for each of the 'twin' feature classes resulted in an identical value for both. It seems plausible that the different spatial index value on the original feature class was a holdover value from its original creation at an earlier time. I would guess that, as the subsequent quantity and spatial qualities of the combined number of features changed, as they had been added to over time, the initial spatial index generated by ArcGIS from that older feature class had resulted in a slightly different value for the new one -- although at the time the second one was created, the points and overall spatial parameters of the two were spatially identical.

Seeing the difference flagged by the table comparison was unsettling, though.

I've attached images and the output of the new table comparison, which now performs exactly as it should:  It identifies that 255 character length text field as the only difference between the new feature class and the original one.

Barry Conner
GIS Analyst
City of Huntsville, Texas

[LauraSilsbee1]

Thanks Barry.  This looks like what I needed.  So far we haven't had any issues but, as you said, it is unsettling when you first see the difference.

Laura