I need to compute the 20 nearest fire stations - out of 51 - from each of the 70,000+ streetlines' midpoints of our county road network.
Since 2007, good ol' Network Analyst for ArcView 3.2 has annually processed such a big amount of calculations thanks to an adhoc Avenue script using 'FindClosestFacilities' function. This script runs like a charm for approx 24 hours to get all done - including reformatting output results into customized tables.
Time has come to switch to up-to-date tools and languages (Python) for hopefully a more efficient processing chain, at least less time-consuming and better integrated with our ArGIS environment.
I gave some tries with ArcGIS' Closest Facility solver, but the OD Cost Matrix is definitely the tool I need, as it is much much faster than the latter and I don't need to output the precise routes feature layer but the tabular data.
The problem is that the matrix I expect requires the 20 fire stations to be set up as the destinations and the 70,000+ centroids as the origins, so that destination ranks range from 1 to 20. But fire engines start from fire stations, right?
The big question is: is it possible, by way of a Python script, to compute not a Origin-Destination Cost Matrix but a Destination-Origin Cost Matrix? with (hidden?) settings like Closest Facility solver's 'TRAVEL_TO' and 'TRAVEL_FROM'?? If not, will the use of ArcGIS' Closest Facility be as reliable and faster than its odl AV3 predecessor?..
Thanks a lot in advance for your help, answers or hints.