Subscribe to this thread
Home - General / All posts - Confused by AOI concepts.
#21-Mar-15 19:43

I have two data sets for at least one, to possibly 13, US State(s). They are polygons features (Areas) that have attached attributes related to Wild Fire potential. One set has values like Low, Moderate, High, and Very High. The other has many classes of urban and residential categories, water & agriculture and Wildland. All data is Lat/Lon. There are several million records in the data sets.

I need to accept a lat/lon pair, or collection of them, and determine the following. 1) Which category the point lands in, for each data-set. 2) The distance, in feet, from the point to the nearest High or Very High category. 3) The distance, in feet, from the point to the nearest Wildland.

I have imported the shapefiles into Manifold and Exported to SQL Server. I’ve set up a project with Linked Drawings using the DatabaseConsole and an Area Of Interest and AR_Rick's AOI add-in and a home grown derivative for the automated approach. I've also cobbled together some queries and control from an external C# application.

It works, but is way slower I would like. Watching Windows Task Manager while the queries run, it seems that Sql Server is active the whole time. It doesn't matter if the queries are run from the GUI or the external app. It seems that Manifold is not using the linked drawing but going back to the server each time.

I thought the whole point of the linked drawing / AOI concept was to work quicker with only the data required. I only need to compare my point to a small number of Area records that are within 5000 ft., not the whole data set. Am I missing something?


bdg

oisink
370 post(s)
#24-Mar-15 18:25

I think the AOI query speeds up the graphical display of data by only fetching the area of interest, but any query on the underlying data will query the whole dataset. (I might be wrong on this!)

But it seems you need to limit your query by using an appropriate filter - see post here http://www.georeference.org/forum/t97659.18#97736 and the entire thread.

Oisin

Sloots

678 post(s)
#24-Mar-15 21:15

A quick test makes clear that a query will use the subset fetched by the AOI, and *not* the whole dataset of the underlying data.


http://www.mppng.nl/manifold/pointlabeler

#31-Mar-15 17:25

My quick test wasn't so quick. These words of wisdom from Adam kind of confirmed it for me. The linked drawing gets refreshed at every query. http://www.georeference.org/forum/t41301.13#41537


bdg

oisink
370 post(s)
#31-Mar-15 18:51

Run it server side and optimize your query, for which I refer you back to http://www.georeference.org/forum/t97659.18#97736

Oisin

#01-Apr-15 12:37

Yeah, I'm kind of grokking the whole filter by by non-spatail component. There is nothing inherit in the data to make that easy. Not looking forward to fabricating intermediate tables, but that seems to be the solution. I'm thinking 5 digit FIPS (US Counties) codes should do it.

Thanks for the input.


bdg

Manifold User Community Use Agreement Copyright (C) 2007-2021 Manifold Software Limited. All rights reserved.