Latest Contributions by JoeBorgione

If I use just end points, doesn't that turn each line in a a street with no turns off of it?  Meaning if I wanted to travel 100 meters east to a parallel street and I was on a north/south street, I would have to travel all the way to the end and turn twice?  I don't think I am explaining that well, but I thought each overlap in this case would be a 4 way intersection so that diagonal lines could be traveled upon. I created the data with your help and if it is correct, then how I solve the problem is the issue.  What I do is add the newly built network dataset to my map when prompted and make sure it is selected in the drop down in the network analyst toolbar.  Then (in the toolbar) I then click network analyst-->new route.  Then I click the create network location tool and add new points which are entered in as stops.  When I click the solve button I get this message every time (and I've tried different locations in the network with different numbers of points and it is always the same): Warning: No route from location "Graphic Pick 1" to location "Graphic Pick 2". Error: No solution found. The concept of turns is you can penalize a route with turns; for example, a left turn may take a little longer than a right turn in a street network.  Or a left turn at 5:00 rush hour is heavily penalized.  As mentioned, turns are an issue for me, but not in your case. So what you are doing now is running a simple point to point route, right?  You're basically testing the connectivity of your fishnet-network, and it sounds like if it's not going to work with two manually entered points, its not going to work with 400.  Any chance you can post a portion of your network here?  I only have a 9.3 install here at home so a shapefile would be fine.  I think you are just missing something in your workflow that I'm not catching. ... View more

-Now it asks me if I want to model turns in this network. I'm not exactly sure what that means in my context. Would the turns be the paths across the polyline fishnet? What should I put here? If yes, are the turn sources just the default global turns? For this project, you don't need to worry about turns. Next menu is connectivity. When I click that and the connectivity groups show up, what options do I take? Do I set the connectivity policy to end point or any vertex? There is another column with a "1" and a check box under it. Do I check that or not? I'm a big fan of of end point only, but I route emergency vehicles to incidents;  if I were to use vertex connectivity I might have an ambulance go from a freeway overpass bridge to a surface street; not good.  In your case, the fishnet should not have any vertices, just endpoints. The next menu asks to specify the attributes for the network dataset. A row with the name Length is there with cost for usage. Do I need to add any parameters or change anything under the evaluators button and menu? For any routing problem you need to minimize some sort value to get the optimum route.  The classic TSP not only minimizes some value, but only allows one stop and one stop only at each of your locations.  Length is all you have to work with so go with it.  In my case, I minimize travel time; I want find the 'closest' unit based on time to get from the station to the incident.  Not a TSP in my case, but  VRP; vehicle routing problem. Of course you'll need to add your stops in order to solve your analysis; be sure that they are all within a reasonable search area when you set up the solve in ArcMap.  Nothing in your post jumps out at me that says you're doing anything incorrectly. Maybe Jay or Patrick with ESRI can take a look here and see what's missing... ... View more

Thanks, I've been trying something like that, but haven't nailed it yet.  The polygon land barrier is needed because these points stretch from Gulf of Maine to the Carolinas and when I did a nearest neighbor route (calculating the closest point, drawing a line to it, then removing it from the list), I would get lines running from Jersey across Cape Cod over land.  And unfortunately, our ship can't fly 🙂 Flying ships probably aren't the answer but if Amazon can use drones....  What's not quite nailed yet? ... View more

Thanks for the response Joe. I was thinking of using network analyst, seemed like the best tool. I have built a simple network data set that seems to work well with the toolbar. for example I can route from points to a destination and so forth. I am still getting results that don't seem to make much sense to me (I am confident this is user error) I ran a closest facility on the network and there is redundancy and lack of sharing in the result. [ATTACH=CONFIG]29628[/ATTACH] Using NA is data intensive; of course that doesn't let the user off the hook....  😉 By data intensive I mean that connectivity, attribute values, etc all have to be up to par or you'll get less than optimal results. ... View more

Hello,  I am looking to solve a seeming simple problem but as this is the first time I have used   spatial analyst professionally its been some time since I have dug into this extension and I can't seem to think though it.      I am trying to solve a simple network. I have a set of locations (green dots) and I am attempting to find the shortest route to connect all dots to the destination (green square) with as much overlap as possible.      As I said before this feels like it would be easy to solve and any help would be appreciated.  [ATTACH=CONFIG]29624[/ATTACH] Are you using spatial analyst or network analyst? If the latter, yes it's a relatively simple problem to solve, but you'll need to first build a network dataset set your parameters for the solver etc. If you truly are using spatial analyst, you might want to post this to the proper forum. Or use network analyst instead.... ... View more

Hi,   I have a feature class with about 400 points that are in the ocean.  I'm trying to use the network analyst extension to find the shortest route to travel to each point (as the crow flies more or less).  I'm having trouble creating my network dataset.  I read in another post that I should make a fishnet feature class (of polylines I'm assuming) of the whole area and use that as the roads.  I would have to add a land feature class as a sort of barrier.  Then I could solve the route with using the reorder option.  None of the tutorials I've read have really shown me how to deal with this situation.  Could someone please point me in the correct direction for solving this traveling salesman problem?  Thanks Don't you hate it when a simple problem isn't really so simple? As mentioned in that other post, use the Create Network tool to create the basis of your network. The extents should be set to include all of your points.  Probably a good idea to use the same projection parameters as your points as well.  Once you get your fishnet built, (you'll be the best judge has to how big the cells should be) put them into a feature dataset and create your network.  You'll have a length attribute to use as your impedance.  I'm not sure what the land feature class barrier idea is for, but it sounds to me like you are ready to go. ... View more

Is there anything wrong with loading data into versioned feature class using the data load function in ArcCatalog? I performed this task on a Street_Centerlines feature class in my own version, but it appears to have caused a massive data corruption to occur in both my version and the Default version.  I hadn't posted the changes, but I did perform the data load outside of an edit session. I tested this process in an inactive sde database, and it worked flawlessly.  When I later attempted it on the active database (which has considerably more overhead), it broke. The goal here was to delete all features from the Street_Centerlines feature class, and load in an updated version of the feature class without disconnecting users.  I know this isn't the best practice, but I was trying to work around the limitations of using Truncate and Append on a versioned data set.  Any thoughts or insights on this topic are appreciated! So the question is why did it work in the test environment and not in the production environment.  The only thing I can think of is that with your users attached, and you deleted the features, that fouled things up.  Not much help I know, but what if you were to boot everyone off and then try it?  My experience with working with SDE is to get everyone off before doing any kind of db mantenance. ... View more

Hi I have a geodatabase feature class that has domains. I have extracted a portion of this feature class and manipulated some of the data within and then appended the data back to the original source. When I append the data the numerical information in the attribute table for the appended portion is lost and replaced by null values.   I have tried to export the portion I need to work on as a shape file and also as a feature class into the same geodatabase. I get the same issue in either case. I was wondering whether the issue with the null values is because the domain for the data fields is lost once I export? If I export the original data and the portion as a shapefile and then merge there isn't an issue. However, I need the domains used in the geodatabase. Any help appreciated. Thanks Sam Domains are a property of the geodatabase, not the featureclass.  When you append the extracted data back into the original database, have you deleted the orginal features?  I ask only because if you have not, you'll end up with double entries.  I'm not a big fan of the append tool.  Must have had a bad experience at one time or another.  Try using the simple data loader.  I also have to ask, why not just perform your manipulations on the original features in the original geodatabase? ... View more

Personally I'd import the spreadsheet into a geodatabase table; that's where I prefer to work.  It looks like your street field has a comma between the actual address and the bldg/apt reference.  With the split() function in VB you can put the bldg/apt reference into it's own field and the actual address in its own field. Add two fields, both text.  Call one TheAddress and the other TheAptBldg Calculate TheAddress: dim list dim addr list = split([STREET],","] addr = list (0) Calculate TheAptBldg dim list dim aptbldg list = split([STREET],","] aptbldg = list (1) Untested.  Your mileage may vary... Also; you might want to look into cleaning up your data:  I go with all caps, and NO punctuation so St. becomes ST etc etc.  Use USPS street type standards.  Don't re-invent the wheel. ... View more

Looking at your data, the only common field that I see is Country in table 2 and ADM0_Name in table 1.  If I am reading table 2 correctly, you have a list of names all in one field (D). Can you try to describe what it is your want to do? ... View more

You're correct - there is no 13 Barwick Road...... but 15 is RIGHT THERE.... I can see it 😉 I'm not as concerned about the 'double ding' for street type and city - we actually prefer that since it makes it more obvious that those candidates are not very similar to what was requested.  The first ding on the house number is a bit of a show stopper though.  It's especially annoying that it worked as desired in 9.3.1 and now it's 'broken' in 10. To me it's working.  Since there is no 13 Barick Rd in Chapple, it can't make a match there.  But it gives you a low scoring partial match elsewhere.  The single address types of locators are just that; single address.  It either hits or it dosen't; there's no range involved.  I'm not second guessing you, it just seems very odd that it worked earlier. If you use a US streets type of locator that provides for a range of addresses you can expect partial hits.  For example if you have an address of 1234 Main St, but the range only goes to 1230, you'll get a partial hit at 1230.  I look at that as suggestion.  It's almost like the locator is saying 'Hey, I don't have exactly what you entered, but I do have this.  Will it do?' ... View more

Yes, The city is correct in both the input address and the data from which the locator was generated. We also use an extensive alias list for city names which is working just fine. Here's a specific example: The source data has a polygon for house_number = 15, street = Barwick, street_type = road, city = TOWNSHIP OF CHAPPLE If you query for 15 Barwick Road, Chapple you get a score of 100, which is correct (first attached image) If you query for 13 Barwick Road, Chapple you get a score of 43, matched to 13 Barwick Drive, Barrie (second attached image) Ideally, I would have gotten a hit on 15 Barwick Road, Chapple with a score close to 100 because it's at least on the right street in the right town and close to the right house number, but it fails to get any kind of score at all.  Instead I get a match (albeit lower score) to the right house number but with a different street type and a totally different city. It seems that for Single House locators, the house number trumps all other address components - meaning it will force matches to the requested house number even if it has to pick the wrong street/town. As far back as I can remember, the single address locator styles are pretty unforgiving. I take it that there is no house number 13 on Barwick Rd in Chapple, right?  If I'm correct, you are getting dinged first because there is no 13 house number in Chapple; after that you get double dinged in the street type and city. ... View more

Phew...  Thought I was having a flash back from the 60s... Again... All seriousness aside: do both the address data and the parcel data have the correct city identifier? ... View more

We have built a locator in ArcGIS 10 SP5 which is based on the US Street - Single Address style. We have been able to tweak most other settings but are still having issues with street number scoring. Basically, we would like the locator to rank similar house numbers on the same street/town with a high score. The default action (from my testing and based on one other forum thread) seems to indicate that the locator is VERY strict about the house number matching - to the point that it will give a high ranking to an address that matches the house number and street in a town hundreds of miles away, but fail to even score the address 100' away. We have tried playing with the scoring of the FullNormalAddress, NormalAddress and House components but nothing seems to allow addresses with close house numbers in the desired city to be ranked higher than exact house number matches in a different city. For example: Search for 100 Main Street, Mytown Geocoder responds with: Score: 90, Address: 100 Main Street, Yourtown (could be large distance away from Mytown and therefore totally wrong location). Yet 98 Main Street, Mytown is only two house numbers away from the desired address.  I would like the locator to return something like: Score: 98, Address: 98 Main Street, MyTown Score: 50, Address: 100 Main Street, Yourtown Has anyone found a solution for this?  Any workarounds?  Apparently this worked as desired in 9.3.1 but was "improved" in 10.0? I'm a little confused with your term "US Street - Single Address " style of locator. I'm familiar with the US Address One range and the US Address Single House.  That latter of which is a point or polygon type of locator. Perhaps you could clarify. see: http://help.arcgis.com/en/arcgisdesktop/10.0/help../index.html#/Commonly_used_address_locator_styles/00250000000v000000/ ... View more

Good day. I am learning Network Analyst. I have noticed, in the tutorial arcgis, that all road split on the intersections. I want to know, must  split road on the intersections for network analyst? Personally, I think it's at least better practice if not best practice to model your roads to be split at intersections.  You have an option in NA to use coincident vertices for connectivity.  But think about it; you could and often times would end up with connectivity between a limited access highway and an over passing or under passing surface street.  Have fun explaining to the boss when you depict a route that jumps or falls 50 feet from one road to another.... ... View more