Calculating distance to median point for a group of points in Model Builder In ArcGIS

Question

I have two shapefiles. First of them, there is about 115 000 points (clients and their field size) All together there is about 16 226 different clients. In the other shapefile I have calculated the median point for all the clients. Now, I want to calculate the distance from each client field to its corresponding median point. How can i do it in Model Builder?? Doing it one by one takes a lot of time.

I really hope someone can help me.

Answer 1

Since you have two shapefiles (one contains median point for each client and another contains source position points), Generate Near Table GP tool might be a good candidate for finding out the distance from each median point to the points around. The only thing left is to iterate over each client's median point without taking into considerations others. We have to select and use only one client per iteration. For me, a Python solution is easier than ModelBuilder based one.

This code should work for 10.0 unless I have missed something. You would need to create a script tool and update paths to the data. You can save the code below in a .py file and then use any text editor to edit the file).

To make it easier to understand the workflow, I've attached a picture. Red points are in one feature class and represent median points, blacks and greens represent source points in another feature class (each have a field ClientID, symbology is based on ClientID, labels are just ObjectIDs).

enter image description here

import arcpy
arcpy.env.workspace = r"C:\GIS\Temp\test.gdb"
arcpy.env.overwriteOutput = True
source_pnts = r"C:\GIS\Temp\test.gdb\source_pnts"
median_pnts = r"C:\GIS\Temp\test.gdb\median_pnts"
iter_neartable = r"NearTableTemp"
sum_neartable = r"NearTableAllClients"
fieldname = "ClientID"
search_cursor = arcpy.SearchCursor(median_pnts,"", "", fieldname)
counter = 0
for row in search_cursor:
 #print row.ClientID
 median_pnts_lyr = arcpy.MakeFeatureLayer_management(median_pnts,"OneClientOnlyMedian","""ClientID = {}""".format(row.getValue(fieldname)))
 print int(arcpy.GetCount_management(median_pnts_lyr).getOutput(0))
 source_pnts_lyr = arcpy.MakeFeatureLayer_management(source_pnts,"OneClientOnlySource","""ClientID = {}""".format(row.getValue(fieldname)))
 print int(arcpy.GetCount_management(source_pnts_lyr).getOutput(0))
 iteration_table = arcpy.GenerateNearTable_analysis(median_pnts_lyr,source_pnts_lyr,iter_neartable, "#","NO_LOCATION","NO_ANGLE","ALL","0","PLANAR")
 if counter == 0:
 arcpy.CopyRows_management(iteration_table,sum_neartable)
 arcpy.TruncateTable_management(sum_neartable)
 arcpy.Append_management(iteration_table,sum_neartable,"NO_TEST")
 counter = counter + 1
 else:
 arcpy.Append_management(iteration_table,sum_neartable,"NO_TEST")

• Kristin: I tried this but it takes account other clients mean centers too, because some are near to each other. I need to classify source points by client number, but no distance tool have case field option to do this.
• Alex Tereshenkov: Ah, I got it. Would you consider an arcpy (Python) based solution? Is it just a one time job?
• Kristin: I never used arcpy solution but yes this is one time job.
• Alex Tereshenkov: I'll publish some code you could run in a while then :)
• Alex Tereshenkov: Start with a test sample data first to see that you get expected results!

Answer 2

I had to come up with a solution for something very similar to this just last week. This is in essence what I did:

Steps

1. Add distance field to your client dataset (the one with 115,000 points)
2. Get the spatial reference of your data
3. Open a search cursor on the mean point, get the X and Y
4. Open an update cursor on the client points. Loop through each point and get the X and Y of the client. Create a table in memory that will hold the xy fields of both the current client and the mean point. Insert the xy data into the table. Note: this table will only have 1 record each time.
5. Create an XY line using the newly created table. This will create a line from the current client to the mean point based on the x and y values, the distance type, and your spatial reference.

6. Open another search cursor on the new xy line, read the shape length and insert it back into the client dataset.

   import arcpy
   def Delete(name):
    if arcpy.Exists(str(name) +'.shp'):
    arcpy.Delete_management(str(name) +'.shp')
    if arcpy.Exists(name):
    arcpy.Delete_management(name)
   def CalculateDistance(mean, clients):
    desc = arcpy.Describe(clients)
    spatialRef = desc.SpatialReference
    with arcpy.da.SearchCursor(mean,['SHAPE@X','SHAPE@Y']) as SC:
    for mean_row in SC:
    mean_x = mean_row[0]
    mean_y = mean_row[1]
    with arcpy.da.UpdateCursor(clients,['SHAPE@X','SHAPE@Y','distance_to_mean']) as UC:
    for client_row in UC:
    client_x = client_row[0]
    client_y = client_row[1]
    mem_table = "in_memory//table"
    Delete(mem_table)
    table = arcpy.CreateTable_management("in_memory","table")
    arcpy.AddField_management(table,"client_x", "DOUBLE")
    arcpy.AddField_management(table,"client_y", "DOUBLE")
    arcpy.AddField_management(table,"mean_x", "DOUBLE")
    arcpy.AddField_management(table,"mean_y", "DOUBLE")
    with arcpy.da.UpdateCursor(table,['client_x','client_y','mean_x','mean_y']) as UC2:
    for tableRow in UC2:
    tableRow[0] = client_x
    tableRow[1] = client_y
    tableRow[2] = mean_x
    tableRow[3] = mean_y
    UC2.updateRow(tableRow)
    mem_line = "in_memory//xy_line"
    Delete(mem_line)
    xy_line = arcpy.XYToLine_management(table,mem_line,'SHAPE@X','SHAPE@Y','mean_x','mean_y',"GEODESIC","ID_FIELD",spatialRef)
    with arcpy.da.SearchCursor(xy_line,["SHAPE@LENGTH"]) as SC2:
    for row in SC2:
    client_row[2] = row[0]
   mean_point = your mean point dataset
   client_points = your client points dataset
   # add distance field to client points
   arcpy.AddField_management(client_points,"distance_to_mean","DOUBLE")
   CalculateDistance(mean_point, client_points)