Here is a little context to the issue. We are mapping occurrences with information coming from two systems the first our Records Management System (RMS), and second the Computer Aided Dispatch System (CAD). The RMS system is updated by our personnel to reflect the address where the occurrence took place, after being dispatch to the location. The CAD system has its information pushed into the RMS system where a coordinate value is updated based on the dispatch location for an occurrence.
In our situation the address having been updated by our personnel, is more accurate than the coordinate, especially considering the occurrence location is not always the same as the dispatch location. For example if a neighbor a couple doors down calls in to report an occurrence, and we used the coordinate coming over from the CAD system the occurrence will be mapped at the neighbors location who placed the call. We want to only use the dispatched location as a secondary option to help fill in the gaps where addresses cannot be geocoded using the address information.
With the improvements to the locators in ArcGIS 10 we have a great opportunity handle this geocode workflow completely within a composite locator, and reduce the processing done using Python and the Geoprocessor. It may also open up additional possibilities such as a setting up a dynamic feature class, and reviewing and rematching our results. This currently is not possible because each geocoding set is processed independently using address locators and xy event layers, with all outputs append together in the end.
