Latest Contributions by RichardFairhurst

Just a word of caution in using the distances generated by the analysis I have proposed (and all of the other suggestions I have seen here).  The distances reported would only really be true and accurate if there was no change in elevation due to either topography or flight ascent and descent.  It is an accurate as-the-bird-flies distance only if the bird were capable of flying nearly at ground level and at a nearly constant elevation. I would tend to assume that adjustments for altitude would support using the shorter distances, because I would assume that the altitude between destinations would not have to be as high in most short distance flights.  However, the distances for take-off and landing or climbs and descents due to terrain and building obstructions would not be accounted for without some kind of elevation data for each point (and possibly elevation data for tall obstruction points that might intervene between your destinations, such as mountains, tall utility obstructions, high-rise buildings, or no-fly restrictions for military bases or other sensitive instalations).  As long as those disclaimers and limitations are noted, the distances obtained from the analysis I proposed are reasonable as an estimate for as-the-bird-flies distances. ... View more

I took my example a step further and after assuming an origin/destination, I was able to manually pick out the shortest non-duplicating path based on a sort of the segment lengths.  Intuitively the result makes sense based on the picture, and if I rearrange any three segments joining three of the points together the total path would have to be longer than the one that is shown.  So even without Network Analyst, this method makes it possible to do the analysis.  Here is a screen shot of the shortest path. I can say the path is 7.99 miles long to make the return trip (after specifying that the starting points was at either the top most or the bottom most point of the set.)  The color based on segment lengths also confirms the path is the shortest.  17 diffferent destinations (with the origin being the final 18th return trip destination) are connected by exactly 17 segments.  The return trip is 3.89 mile ling, which under half of the total trip length.  The average length of the segments is .47 miles long and the shortest segment is 216 feet or 0.04 miles long. ... View more

Actually not, that will not be problem. This is short distance flying with helicopters, in a limited area with many locations spread out. As the distance is so short, a linear distance in the projected map is just fine. My big problem is how to calculate and present the order that will result in the shortest total distance I don't deal with flight paths, but this sounded like an interesting problem that expands the potential interplay of intermodal transportation networks, which is a problem transportation analysts like myself are supposed to tackle. I beleiev you can convert your potential flight paths into a network that Network Analyst can use to do the analysis.  Here is how you can create the network and eliminate duplicate reversed lines (so each segment can function as a two-way street in effect).  You already have all of the points and they all identify the helicopter associated with them.  So here is what I would do with them: 1.  Make sure the point data is in a geodatabase.  I prefer a file geodatabase for precision and processing speed.  2.  Add an integer field to your point data named ORIG_ID and calculate the ObjectID values of the points into it. 3.  If your points are using a Geographic Coordinate System (i.e, using coordinates in degrees like Longitute and Latitude), then create a reprojected copy into a Projected Coordinate System (using Linear Coordinates, like a State Plane projection).  This will allow you to calculate linear distances between points, since you are not needing to account for geodesic distances. 4.  Add two double fields for your X and Y coordinates and use the Geometry Calculator to calculate the projected point coordnate values into the fields. 5. Make a copy of the projected coordinate system point data into the same geodatabase with a slightly different name. 6. Use Make Query Table to do a many to many join of the point data to its duplicate.  You do not want include the shape field so that the output is just a Standalone table.  To create all of the point pairs for a given helicopter set the query criteria to be something like this:    orig.HelicopterID = copy.HelicopterID 7.  Select from the Query Table all records where orig_OBJECTID < copy_OBJECTID to get all paths that will not have a reversed duplicate or from the point to itself.  Export the results of that selection to a Standalone Table. 8.  Add a field called PATHID and calculate the OBJECTID of the exported matched output into it. 9.  Use the Make X/Y event Layer with the X and Y coordinates from the Original X/Y coordinates.  Export that point set. 10. Use the Make X/Y event Layer with the copy X and Y coordinates.  Export that point set. 11.  Merge the two points sets together. 12.  Use the Points to Line tool to create segments between all points that will not be duplicated (assuming you had no overlapping points to begin with) using the PATHID value as the LINE ID of the tool and the sequence as the ObjectID of the merged points. 13.  Add a field to the paths for HelicopterID to the paths.  Join the Paths back to one of the unmerged Point outputs using the PATHID as the join field and calculate over the associated HelicopterIDs. Now you can select all of the paths possible for the helicopter to travel and will have the length field to tell you how long each segment is.  You should also be able to transfer any other information you want that came from the original points that tie to each path using this join. The result would look something like the attached screenshot (created from all node points along a roadway as a demonstration of the concept).  The colors of the segments reflect their relative lengths and just one set of segments is shown related to a single roadway (like showing a single helicopter).  These are all of the possible straight line as the bird flies paths between nodes. Network Analyst can most likely be set up with constraints that would keep the helicopter on its possible flight paths and also account for each point accessible along the possible path to solve for the shortest distance.  Constriants such as specifying an origin point and whether to make the end point a return to the origin point should also be possible. I hope this helps. ... View more

Apparently this behavior is not limited to using an Add-In or a Windows Form.  The mere act of starting and stopping Customize Mode without doing anything else causes the frozen Layout behavior.  Or just pressing a user created Add-In button on a toolbar that does absolutely nothing causes the behavior.  (Buttons created by ESRI do not seem to have this problem, just user Add-In buttons). At this point it seems clear to me that this is a fundamental bug in the Desktop application affecting all user customizations and that it is not tied to any particular kind of user customization.  I would consider this to be a serious issue that indicates that the overhaul of user customizations that occured in 10.0+ was poorly implemented compared to previous versions. ... View more

David: I have modified your replacement DataStatistics code slightly. I have added a new property that returns an UniqueValueCount so that the class supports the use of a For loop in the same way as IDataStatistics. Your Interface declarations were misspelled. It read IDataStatiticsECL and not IDataStatisticsECL. That error made it harder to incorporate into my code that used the IDataStatistics interface. The spelling correction below makes it easy to just append "ECL" to the appropriate places in code that uses the broken IDataStatistics interface. This fixed the crashes and memory errors I was having with the IDataStatistics interface, so I appreciate the code you posted. It saved me a lot of time. Overall the performance of the new code seem fairly good, but if any one has performance enhancement suggestions, I am also intersested. Here is the revised code: Imports ESRI.ArcGIS.Geodatabase Public Interface IDataStatisticsECL     WriteOnly Property Cursor() As ESRI.ArcGIS.Geodatabase.ICursor     Property Field() As String     ReadOnly Property UniqueValues() As System.Collections.IEnumerator     ReadOnly Property UniqueValueCount() As Integer End Interface Public Class DataStatisticsClassECL     Implements IDataStatisticsECL     Private m_Cur As ESRI.ArcGIS.Geodatabase.ICursor     Public WriteOnly Property Cursor As ESRI.ArcGIS.Geodatabase.ICursor Implements IDataStatisticsECL.Cursor         Set(ByVal value As ESRI.ArcGIS.Geodatabase.ICursor)             m_Cur = value         End Set     End Property     Private m_Field As String = ""     Public Property Field As String Implements IDataStatisticsECL.Field         Get             Return m_Field         End Get         Set(ByVal value As String)             m_Field = value         End Set     End Property     Public ReadOnly Property UniqueValues As System.Collections.IEnumerator Implements IDataStatisticsECL.UniqueValues         Get             Dim row As Row = CType(m_Cur.NextRow(), Row)             Dim dic As New SortedList(Of Object, String)             Dim iFeild = row.Fields.FindField(m_Field)             'loop through the data and get the unique values             Do Until row Is Nothing                 If dic.ContainsKey(row.Value(iFeild)) = False Then                     dic.Add(row.Value(iFeild), "")                 End If                 row = CType(m_Cur.NextRow(), Row)             Loop             Dim arr(dic.Count - 1) As Object             Dim arrInd As Integer             dic.Keys.CopyTo(arr, arrInd)             m_UniqueValueCount = arr.Length             Dim n As New ValuesEnum(arr)             Return n         End Get     End Property     Private m_UniqueValueCount As Integer = 0     Public ReadOnly Property UniqueValueCount As Integer Implements IDataStatisticsECL.UniqueValueCount         Get             Return m_UniqueValueCount         End Get     End Property     Private Class ValuesEnum         Implements IEnumerator         Private ReadOnly m_Values() As Object         Private m_Cur As Integer         Public Sub New(ByVal lst() As Object)             m_Values = lst             m_Cur = -1         End Sub         Public ReadOnly Property Current() Implements IEnumerator.Current             Get                 If m_Cur >= m_Values.Length Then                     Return Nothing                 End If                 Return m_Values(m_Cur)             End Get         End Property         Public Sub Reset() Implements IEnumerator.Reset             m_Cur = -1         End Sub         Public Function Movenext() As Boolean Implements IEnumerator.MoveNext             m_Cur = m_Cur + 1             If m_Cur < m_Values.Length Then                 Return True             Else                 Return False             End If         End Function     End Class End Class ... View more

Richard, Did you see this thread that discovered the bug NIM087476  �??Memory leakage ESRI.ArcGIS.Geodatabase.IDataStatistics using UniqueValues property  with ArcGIS 10.1 with SP1�?� Thanks for pointing that out.  I had not seeen this thread, but this explains why it worked at 10.0 and not at 10.1 SP1 (I skipped 10.1 original).  I will reengineer the code to use a cursor to get unique values.  Can anyone pioint me to any threads that show the most elegant techniques for doing that? ... View more

I wrote an Add-In that used a Windows form and many chained functions that create and release cursors using ReleaseComObject. Two of the functions use IDataStatistics, which takes a populated cursor as a parameter. When I was targeting ArcGIS 10.0 it all worked fine. Now with 10.1 I am getting a lot of errors about trying to access protected memory, which the MSDN help suggests is due to memory corruption, usually related to unmanaged code. The error help suggests that the memory corruption may be accumulating through many operations before triggering the error. The failure seems to always happen on the second or third pass that is triggered by user form updates and occurs within the subroutines that have the cursor that feeds to an IDataStatistics object. The error either appears when I call the UniqueValues method of the IDataStatistics object or when I do a Search on the feature cursor before it is passed to the IDataStatistics object (only on the second or third time through the code, never the first). Its behavior can be altered by the use or lack of use of the ReleaseComObject statements for the cursor or the IDataStatistics. I have tried all of the combinations I can think of and orders of the releaser. Below is a sample of how the code is written with the two lines where the protected memory error can occur shown in Red. Any suggestions on what objects I should use the ReleaseComObject on and the order or placement of that code or how to trap these errors and issolate the real cause?     Private Sub GetRoutesList()         ' Set Up Query Filter for the Full Date Range of the Collision Layer respecting PDO and DUI flags         Dim pQF As IQueryFilter = New QueryFilterClass         pQF.WhereClause = "NOT " & fldCollisionDate & " IS NULL" & whereSeverity & whereDUI         pQF.SubFields = fldCollisionDate & ", " & fldPrimaryRoad & ", " & fldRID & ", " & fldSeverity & ", party_sobriety_1, party_sobriety_2"         ' Query Collisions Date Range and Street to fill Route Combobox List         pQF.WhereClause = fldCollisionDate & " >= date '" & m_FromDate.ToString("yyyy-MM-dd HH:mm:ss") & "' AND " & fldCollisionDate & " <= date '" & m_ToDate.ToString("yyyy-MM-dd HH:mm:ss") & "' AND " & fldRID & " > ' ' AND Upper(" & fldPrimaryRoad & ") = '" & m_strPrimary & "'" & whereSeverity & whereDUI         Dim dataStatisticsRoutes As IDataStatistics = New DataStatisticsClass         Dim featCursorRouteList As IFeatureCursor = m_FLCollisions.Search(pQF, True)         ' Use DataStatistics to get Unique Values         dataStatisticsRoutes.Field = fldRID         dataStatisticsRoutes.Cursor = featCursorRouteList         Dim pEnum As System.Collections.IEnumerator = dataStatisticsRoutes.UniqueValues()         ' Fill the Combobox and suspend refresh until list is populated         cboRoutes.BeginUpdate()         ' Clear the Routes Combobox         cboRoutes.Items.Clear()         ' Reset the enumerator before trying to access its values         pEnum.Reset()         ' Read each Unique value and populate the Route list.  Sorting is handled by a Combobox property.         For i As Int32 = 0 To dataStatisticsRoutes.UniqueValueCount - 1             pEnum.MoveNext()             cboRoutes.Items.Add(UCase(pEnum.Current()))         Next i         ' Release the QueryFilter         If Not pQF Is Nothing Then             System.Runtime.InteropServices.Marshal.ReleaseComObject(pQF)         End If         ' Release the dataStatistics         If Not dataStatisticsRoutes Is Nothing Then             System.Runtime.InteropServices.Marshal.ReleaseComObject(dataStatisticsRoutes)         End If         ' Release the Cursor         If Not featCursorRouteList Is Nothing Then             System.Runtime.InteropServices.Marshal.ReleaseComObject(featCursorRouteList)         End If         ' Allow the Combobox to refresh         cboRoutes.EndUpdate()         If cboRoutes.Items.Contains(m_strRoute) Then             ' Set Combobox Text to the current Map Route             cboRoutes.Text = m_strRoute         ElseIf cboRoutes.Items.Count > 0 Then             ' Set Combobox Text to First Route in the List             cboRoutes.Text = cboRoutes.Items(0)             m_strRoute = cboRoutes.Text         End If     End Sub There are many other cursor calls that do not involve IDataStatistics and other Com objects from ArcGIS being created in the code. None of the cursors are embedded in each other and I have always used the ReleaseComObject call after I finish reading the cursor. These cursors have not generated errors so far, but I need to be sure they are not contributing to the problem. Any suggestions are appreciated. ... View more

This is a screenshot I want to inquiry to know the distance betweeen ex: 119-107 (Shortest road) 172-136 148-117 .. etc. Blue Points are in a Shape File Road (Red & Green) are in a Shape file. [ATTACH=CONFIG]20992[/ATTACH] You are describing a Network Analyst problem.  You need that extension to trace road segments between two destinations along the shortest path.  You also have to create a Network Dataset from your road network to use the tool.  Look in the help for Network Analyst and then you may want to go to that Forum topic to ask more questions. ... View more

Hello, I am fairly new in ArcGIS. I am wondering if there is a short-cut way of connecting closely spaced lines. I mean I have hundreds of lines (pathway for example), but some of the lines are not connected (small gaps exist between some lines). How can I connect those? I know I can manually connect by extending one line or drawing a new one in between, but it would require lot of time as I have hundreds. If you know a simple way of doing this (connect line if the distance in between xxx), I will be pleased to know. Thanks Shafayat If you have an Standard (ArcEditor) or Advanced (ArcInfo) licence you can use the Extend Line tool to do the whole feature class at once.  If you have a Basic (ArcView) license then I am not sure what 3rd party tools or scripts are out there.  But hopefully if you are maintaining data you really should get access to more than a Basic license (at least within your organization). ... View more

I've been using ESRI products since ARC/INFO 3.  I recall being able to perform a join or setup relationships between tables using different techniques.  For example indicating the data in the related column is sorted and thus you could use the ORDERED option.  This naturally sped up the join or relate a great deal. I'm using the following python code... arcpy.JoinField_management(pointFC, "joinfld_name", pointattributes, "joinfld_name") to join a table of 200,000 records to a point FC of exactly the same number of records (i.e. 1:1 match). This join takes more than 25 minutes.  I suspect it takes so long because it's not performing any intelligent searching on the join column of the "pointAttributes" table.  Instead it's just doing a brute force search. I've tried creating an attribute index on both columns but this didn't seem to help. All of the data is in the same fGDB. I'm running Windows 7 with ARcGIS 10.0 on a relatively new laptop.  Thanks for any help. Regards, Brian JoinField_Management is permanantly and directly physically altering the input table and writing it to disk.  That will take a long time, indexed or not.  If you are only wanting a join like you would create through the context menus in the Desktop and expecting that kind of performance, then you need to use AddJoin_management, not JoinField_Management.  The AddJoin_Management tool creates an in memory join that is not written to disk unless you copy or export it.  In memory processes are many orders of magnitude faster than any process that writes to disk.  I hope this helps. ... View more

Hi there, I am building an optimization tool, Just wondering if anyone knows how to: Firstly, export classified values into separate feature classes (i.e. I classify points using symbology, now I want to export those classes as separate shapefiles).. without having to go in and manually select by attribute. Secondly, Use an iteration, either in modelbuilder or python to iterate the break values of the aforementioned classified feature class.  I think this is tricky because it requires the manipulation of the symbology window, which I'm not sure is possible in modelbuilder. Some background / context: What I'm looking to do is classify points by angle, into 5 different quantile classes.  Export those, connect them using least distance, evaluate them based on landuse and water / pipe crossings (give the path a value based on its desirability) then go back to the point shapefile and reclassify, slightly shifting the break values in the quantile classification until all possible sequences are found and evaluated.  The tool will then choose the lowest value as the most desirable and ta da we have an optimized layout! Any ideas would be much appreciated, if my question doesn't make any sense please let me know and I'll clarify!! Thanks! Make it easier on your self.  Create a new long or text field to hold a value representing your 5 different quantile classes and create a single Code Blocked field calculator expression with if clauses that test each record for which of the five classes the value occurs in.  You may want to add a test for Null values as well to avoid an error when that occurs.  Now with that field you should have options to break apart separate feature classes in a variety of ways. ... View more

Yes, that sounds perfect.  How do I run summarize attributes summary? Use the Dissolve tool under the Generalize group of the Data Management group to create multipart lines that will contain the lengths of the combined lines.  The Summary Statistics tool under the Analysis geoprocessing tools can also be used to get a table output, but since you are wanting two case fields (Type and Line ID), it is hard to join back the results to the original lines, whereas the Dissolve tool will give a spatial output that can be joined back to the original lines using a Spatial Join. ... View more