GIS Life Blog

In order to support a very large number of map consumers, that don't ever edit anything or create data, I rely on AGOL to host Map (Hosted Feature Services), hook those back to Portal as "Items", and keep my ArcGIS Server Farm service instances low, only spawning what is needed to support editors. Through security controls, the map service that automatically comes with a feature services is un-discoverable. I don't want people hitting the map services and spawning more service instances. This is in effect a "Collaboration"for people that can't enable Collaboration on their Portal. 

https://www.esri.com/arcgis-blog/products/api-python/analytics/updating-your-hosted-feature-services... has been pretty handy for the past few years, via a scheduled task, it updates (nightly) several hundred feature services by pulling data from SDE into a Pro Project and overwriting what's on AGOL. Unfortunately, it doesn't allow you to preserve "Allow Export to other formats" setting on AGOL. 

Enter https://pro.arcgis.com/en/pro-app/arcpy/sharing/introduction-to-arcpy-sharing.htm, which I've finally rolled my sleeves up on and have converted all my python doo-hickeys to use https://pro.arcgis.com/en/pro-app/arcpy/sharing/featuresharingdraft-class.htm, https://pro.arcgis.com/en/pro-app/tool-reference/server/stage-service.htm, and https://pro.arcgis.com/en/pro-app/tool-reference/server/upload-service-definition.htm. Coincidentally there's a lot more options to control service parameters. 

Disclaimer: I know nothing about python, so I'm sure there's all sorts of inefficiencies in here, but it works. 

import arcpy
import sys, string, os, calendar, datetime, traceback,smtplib
from arcpy import env
from subprocess import call

# Mail Server Settings
service = "GRSM_SASQUATCH"
sd_filename = service + ".sd"
try:
    d = datetime.datetime.now()
    log = open("C:\\PYTHON_LOGS\LOG."+service+".txt","a")
    log.write("----------------------------" + "\n")
    log.write("----------------------------" + "\n")
    log.write("Log: " + str(d) + "\n")
    log.write("\n")
# Start process...
    starttime = datetime.datetime.now()
    log.write("Begin process:\n")
    log.write("     Process started at " + str(starttime) + "\n")
    log.write("\n")
    
    # Mail Server Settings
    SERVER = "1.2.34"
    PORT = "25"
    FROM = "sasquatch@big.foot.com"
    MAILDOMAIN = '@big.foot.com'
    # Data Steward getting the email. Needs to be their email address...without @big.foot.comat the end
    userList=["yeti"]
    # get a list of usernames from the list of named tuples returned from ListUsers
    userNames = [u for u in userList]
    # take the userNames list and make email addresses by appending the appropriate suffix.
    emailList = [name +  MAILDOMAIN for name in userNames]
    TO = emailList
    # Grab date for the email
    DATE = d

    # Sign in to portal
    arcpy.SignInToPortal('https://www.arcgis.com', 'userid', 'password')

    # Set output file names
    outdir = r"C:\PRODUCTION\GRSM_SASQUATCH"
    sddraft_filename = service + ".sddraft"
    sddraft_output_filename = os.path.join(outdir, sddraft_filename)
    #Delete any left over SD files from failed previous run
    try:
        os.remove(sd_filename)
        print("Successfully deleted ", sd_filename)
    except:
        print("Error while deleting file ", sd_filename, ", perhaps it doesn't exist")
    try:
        os.remove(sddraft_output_filename)
        print("Successfully deleted ", sddraft_output_filename)
    except:
        print("Error while deleting file ", sddraft_output_filename, ", perhaps it doesn't exist")        

    # Reference map to publish
    aprx = arcpy.mp.ArcGISProject(r"C:\PRODUCTION\GRSM_SASQUATCH\GRSM_SASQUATCH.aprx")
    m = aprx.listMaps("GRSM_SASQUATCH_LOCATIONS")[0]

    # Create FeatureSharingDraft and set service properties
    # https://pro.arcgis.com/en/pro-app/arcpy/sharing/featuresharingdraft-class.htm
    sharing_draft = m.getWebLayerSharingDraft("HOSTING_SERVER", "FEATURE", service)
    sharing_draft.summary = "Sasquatch Locations"
    sharing_draft.tags = "Sasquatch, Fur, Hairy, Big Foot"
    sharing_draft.description = "Hide and Seek Champion"
    sharing_draft.credits = "Yeti"
    sharing_draft.useLimitations = "This is not real"
    #sharing_draft.portalFolder = "Front Country"
    sharing_draft.overwriteExistingService = "true"
    sharing_draft.allowExporting = "true"

    # Create Service Definition Draft file
    sharing_draft.exportToSDDraft(sddraft_output_filename)

    # Stage Service
    # https://pro.arcgis.com/en/pro-app/tool-reference/server/stage-service.htm
    sd_output_filename = os.path.join(outdir, sd_filename)
    arcpy.StageService_server(sddraft_output_filename, sd_output_filename)

    # Share to portal
    # https://pro.arcgis.com/en/pro-app/tool-reference/server/upload-service-definition.htm
    print("Uploading Service Definition...")
    arcpy.UploadServiceDefinition_server(sd_output_filename,
                                         "My Hosted Services",
                                         "",
                                         "",
                                         "EXISTING",
                                         "existingFolder",
                                         "",
                                         "OVERRIDE_DEFINITION",
                                         "SHARE_ONLINE",
                                         "PUBLIC",
                                         "SHARE_ORGANIZATION",
                                         ["GRSM","Great Smoky Mountains National Park Open Data"] )

    # Clean up SD files
    try:
        os.remove(sd_filename)
        print("Successfully deleted ", sd_filename)
    except:
        print("Error while deleting file ", sd_filename, ", perhaps it doesn't exist")
    try:
        os.remove(sddraft_output_filename)
        print("Successfully deleted ", sddraft_output_filename)
    except:
        print("Error while deleting file ", sddraft_output_filename, ", perhaps it doesn't exist")        

    # Write nothing to log if success.
    endtime = datetime.datetime.now()
    log.write("     Completed successfully in " 
           + str(endtime - starttime) + "\n")
    log.write("\n")
    log.close()
    print('done')

except:
    
 # Get the traceback object
    tb = sys.exc_info()[2]
    tbinfo = traceback.format_tb(tb)[0]
 # Concatenate information together concerning 
 # the error into a message string
    pymsg = "PYTHON ERRORS:\nTraceback info:\n" + tbinfo + "\nError Info:\n" + str(sys.exc_info()[1])
    msgs = "ArcPy ERRORS:\n" + arcpy.GetMessages(2) + "\n"
# Return python error messages for use in 
# script tool or Python Window
    arcpy.AddError(pymsg)
    arcpy.AddError(msgs)
# Print Python error messages for use in 
# Python / Python Window
    log.write("" + pymsg + "\n")
    log.write("" + msgs + "")
    log.close()
    # Define email message if something went wrong
    SUBJECT = "Notification of Un-Successful AGOL Update of "+service 
    MSG = "Did Not Update: {} - ID: {} at "+ str(DATE)+ "; " +pymsg + "; " + msgs
    print (MSG)
    print (emailList)
 
    # Send an email notifying steward of successful archive
    #MESSAGE = "\ From: %s To: %s Subject: %s %s" % (FROM, ", ".join(TO), SUBJECT, MSG)
    MESSAGE = "Subject: %s\n\n%s" % (SUBJECT, MSG)
    try:
            try:
                print("Connecting to Server...")
                server = smtplib.SMTP(SERVER,PORT)
                try:
                    print("Login...")
                    try:
                        print("Sending mail...")
                        server.sendmail(FROM, TO, MESSAGE)
                    except Exception as e:
                        print("Send Error Mail\n" + e.message)
                except Exception as e:
                    print("Error Authentication Server: check the credentials \n" + e.message)
            except Exception as e:
                print("Error Connecting to Server : check the URL of the server and communications port ( 25 and ' the default ) \n" + e.message)
     
            print("Quit.")
            server.quit()
     
    except Exception as e:
            print (e.message)    
‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍

Many of us have one or two things about Pro that is a "fish or cut bait" barrier to migrating to Pro. For me, it's ArcGIS Pro Publish Services To ArcGIS Server. My thoughts, and those of many others, on that topic are well and clearly laid out in that idea, so no rehashing the debate here. But what's really frustrating the very many folks that are impacted by not being able to perform that critical task is the lack of any meaningful response to the clearly articulated concerns, use cases, needs for the requirement, and impact of not having it that have been repeated in the over 150 comments provided. After a "few beers" last night, a conspiracy theory developed that involved geography, learning how to symbolize things in Pro, and the top 100 Ideas by number of votes. Read on....

In the Kingdom of Sasquatchia, there are 100 states and the Constitution's only Amendment "Thee Shall Not Conform to the Agile Process" has survived numerous Supreme Court challenges and two referendums. A bill to codify "Allow ArcGIS Pro To Publish to Server Without Requiring Portal" has broad bipartisan support but not enough votes to override what the President (Big Foot) has been Tweeting "Is a terrible idea! Terrible! Bad Deal! Veto!!!". Concerned about chances of reelection and a trade war with the Boundless Union of Republics, Congress turned to the same pollsters that predicted the outcome of the most recent election. 

Their conclusion: There's no way to target campaign ads based on the geographic distribution of states that are successful in getting ideas adopted. Red and orange states are randomly scattered and green-ish states offer no clue based on their location as to why ideas languish or are not considered. The statisticians where consulted: 

When mapping the number of Vote normalized by number of Views we clearly see very few red states achieved success with a massive number of votes (over 100000) but then a some of the orange states achieved success with 3000 or less votes. Different statisticians were consulted: 

Which substantiates the theory that "Greater number of votes correlates to an idea being implemented". Cartographers were consulted: 

And they produced a map showing just those states with a status of "New"

Compared to all others states.

The Pro Chart Wizard was consulted:

And the campaign strategy manifests itself! Ideas still considered New also have a very high number of views but not a lot of votes when compared to Implemented Ideas with fewer views. 

I often here people say "Don't complain about politics if you don't vote". There's a lot of eyes on ArcGIS Pro Publish Services To ArcGIS Server yet, despite having the greatest number of views and comments of any of the top 100 Ideas (by number of votes), it has the least likely chance of becoming implemented, even less so than Google Maps Basemap . 

Get out the Vote! 

Let's make ArcGIS Pro Publish Services To ArcGIS Server the number one voted Idea on Geonet. Vote with your keyboard, not your mouse, and be sure to include your use case with your upvote! As Pro shows, ideas with a lot of votes get implemented! 

And if you concluded that not a single statistic in the maps show makes one iota of sense, just remember, Arc GIS Pro, not Russians, can swing the outcome of an election.....

Data used to create these wonderful maps in Pro is located at https://bigfoot.maps.arcgis.com/home/item.html?id=8c72924d814e434b8a5fcb08f29ba758 

Have you ever gotten this Error: The selected field is a required field and cannot be deleted ? Scratching my head, wondering how I checked "Field IsRequired". The official solution, export the feature class to another feature class and do some field mapping magic, then delete the original, import the changed one......not. See, all of my feature classes are baked into an enterprise geodatabase with all kinds of dependencies and relationships (that's kinda the point of an enterprise geodatabase, right?). 

There's a quicker and easier way to resolve this issue. I spend a lot of time editing metadata in its native XML (faster, and I can automate things easier), so why not edit a feature class definition, which is also in XML? Feature class definitions live in a XML column (Definition) in the GDB_ITEMS table. 

Click on the xml in the definition column, and you'll see the raw schema for the feature class. Look for the name of the attribute that allegedly won't let you delete it: 

<GPFieldInfoEx xsi:type="typens:GPFieldInfoEx">
 <Name>GIS_LOCATION_ID</Name>
 <ModelName>GIS_LOCATION_ID</ModelName>
 <FieldType>esriFieldTypeGUID</FieldType>
 <IsNullable>false</IsNullable>
 <Required>true</Required>
 <Editable>false</Editable>
 </GPFieldInfoEx>‍‍‍‍‍‍‍‍

Note the Required tag. Ok, so we've figured out what's causing the problem, now, how to fix it. Updating  XML in SQL is pretty straightforward, if you're replacing the entire string. Updating specific nodes and elements in SQL is a little trickier (but possible). In the code below, note the =N'  before the xml definition. I've left out a lot of the other xml, but what I did was copied the entire XML definition and pasted it after the =N' . There's a ' at the end. Also note the WHERE [Name] = 'WETLANDS.DBO.GRSM_HYDRO_WETLAND_PT' which constrains the update to just that feature class with the supposedly undeletable attribute.

UPDATE [dbo].[GDB_ITEMS]
SET [Definition] =N'<DEFeatureClassInfo xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns:typens="http://www.esri.com/schemas/ArcGIS/10.5" xsi:type="typens:DEFeatureClassInfo">
 <CatalogPath>\WETLANDS.DBO.GRSM_HYDRO_WETLAND_PT</CatalogPath>
 <Name>WETLANDS.DBO.GRSM_HYDRO_WETLAND_PT</Name>
 <ChildrenExpanded>false</ChildrenExpanded>
 <DatasetType>esriDTFeatureClass</DatasetType>
 <DSID>435</DSID>
 <Versioned>false</Versioned>
 <CanVersion>true</CanVersion>
 <ConfigurationKeyword />
 <RequiredGeodatabaseClientVersion>10.2</RequiredGeodatabaseClientVersion>
 <HasOID>true</HasOID>
 <OIDFieldName>OBJECTID</OIDFieldName>


blah blah blah

 <GPFieldInfoEx xsi:type="typens:GPFieldInfoEx">
 <Name>GIS_LOCATION_ID</Name>
 <ModelName>GIS_LOCATION_ID</ModelName>
 <FieldType>esriFieldTypeGUID</FieldType>
 <IsNullable>true</IsNullable>
 </GPFieldInfoEx>

blah blah blah

 <ChangeTracked>true</ChangeTracked>
 <FieldFilteringEnabled>false</FieldFilteringEnabled>
 <FilteredFieldNames xsi:type="typens:Names" />
</DEFeatureClassInfo>'
WHERE [Name] = 'WETLANDS.DBO.GRSM_HYDRO_WETLAND_PT'‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍

You'll see, somewhere in the middle, I changed the attribute restrictions to be just like any other attribute: one I can delete. 

Close and restart Arc Catalog to see the change (and delete the attribute). It goes without saying that you should pull a full back up of your SQL before doing this, and should test this on a non-production database, first! Also note this won't work on a GlobalID field (e.g if you have a GLobalID attribute but it's not named GlobalID), there's a few more changes you have to make to alter that one, and you shouldn't view this method as a way to alter the table schema physical design. 

Disclaimer: Editing the underlying GDB tables in this manner is not not supported by ESRI..with that out of the way, there's a lot of DB administration you can pull off with manipulation of the XML definitions. I've used this with great success in exporting a "workspace", editing the XML to implement some find and replace changes, then import as a new GDB. 

This is really just to help me "bookmark" all of the SSL issues, as we're having many of them, hope it helps other folks out!

1. Secure SSL services in Web App builder failing.
2. ArcGIS Portal 10.3 - SSL Certificate for internal network QUESTIONS
   1. Includes a little-known REST checker that test certificate status: cool! https://%3Cserver.domain.com%3E:7443/arcgis/sharing/checkUrl.jsp?url=%3CinsertURLthatisboundtothecer...
3. ArcGIS Portal 10.3 - SSL Certificate for internal network QUESTIONS
   1. Poster can't seem to get a cert working, explored using their own CA server
4. Setting up SSL with a signed certificate in a multi-tier server architecture
   1. Take home message: use proper DNS Aliasing!
5. SSL and esri/request
   1. Seems that enabling "HTTPS" only causes some app errors downstream
6. Anyone else using SSL with AWS EC2 and elastic load balancer?
   1. Don't use Amazon ECW if you need PKI!
7. Sharing Web GIS Services? Always enable TLS | ArcGIS Blog
   1. Another test of SSL functionality SSL Server Test (Powered by Qualys SSL Labs)
8. how to use portalpy with win-authenticated SSL
   1. PKI- Portal doesn't work with python
9. ssl certificate to publish the geoprocessing tool
   1. Enabling SSL breaks GP Tools
10. ArcGIS Server with Reverse Proxy using IIS and SSL
    1. The horrors of reverse proxy
       1. 40694 - Set up a reverse proxy with ArcGIS 10.1 for Server on IIS ARR with SSL
11. How many SSL certificates does ArcGIS Server need?
    1. Certs must use FQDN
12. Issues enabling SSL and accessing with HTTPS
    1. Another confirmation that wild card certs don't work
13. Is it necessary to enable SSL if the ArcGIS Web Adaptor forwards request 2 ServerSite
    1. SSL is required to encrypt passwords
14. Creating SSL Certificates for ArcGIS Server - do i need to use AGS Admin Interface
15. CA Signed SSL Installation with Web Adpater on IIS 7.5
    1. remember, ArcGIS Server is a separate Apache Web Server: Needs a trusted cert as well! AGIS Cert needs to be signed by trusted CA
16. Latency on get token request via SSL (ArgGIS server 10.1 SP1)
    1. SSL kills performace. What a surprise.....
17. Wildcard SSL installation help!
18. Adding unsecured services
    1. This is a big problem: If organization requires HTTPS only, can't consume HTTP content in portal
19. PKI auth. In portal with AD
20. How to connect to a secure ArcGIS Server using HTTPS and a PKI (X509) Certificate
21. Federating an ArcGIS Server site with your portal

There are about as many opinions and resources on SQL Spatial Index Tuning as there are on why the use of a Sequential GUID is good or bad. There's The Black Art Of Spatial Index Tuning In SQL Server | boomphisto , ArcGIS Help 10.1, Spatial Indexing Overview , Spatial Indexing: From 4 Days to 4 Hours - CSS SQL Server Engineers - Site Home - MSDN Blogs, sp_help_spatial_geography_histogram and Indexing geography data in SQL Server Denali | Alastair Aitc... , Basic Multi-Level Grids - Isaac @ MSDN - Site Home - MSDN Blogs , http://social.technet.microsoft.com/wiki/contents/articles/9694.tuning-spatial-point-data-queries-in... ,  and of course, if you've ever struggled with a poorly-performing spatial index and posted about it on the internet, chances are you got a reply from this guy.

Unfortunately, most of us are not database programmers, nor do we dream in SQL syntax. SQL is like learning French by reading a Chinese dictionary. I don't know about you, but I can't make heads or tails of one single line of SQL code. I'm happy to copy other peoples SQL snippets and hack them up 'till I get them to work. I particularly love this statement from ArcGIS Help: "If you create your data through ArcGIS for Desktop, the spatial grid index is calculated for you." That's like saying "If you put your key in the ignition, your car will drive to the store, get milk, and NOT kill a raccoon on the way". Unless your a Massachusetts driver, there are obviously a few things you have to tweak, such as turn the steering wheel and press the gas paddle in endlessly unique combinations in order to get to the milk store and swerve to avoid the raccoon. The same goes for SQL Spatial Indexes in a SDE Database. Just because the software enables them out of the box by no means implies that the index is optimized for your particular data environment.

Out of the box, ArcGIS creates a spatial index with 16 Cells Per Object with all four levels set to Medium Grid Levels.

Notice how the storage type is Goemetry, which some of you are discovering is the "New" ESRI default storage format.

Let's see how the "Out of the Box" spatial index performs. One way to test a spatial index "Quick and Easy" is with the Spatial Index Stored Procedures.

CREATE SPATIAL INDEX 
--THIS IS THE DEFAULT NAME OF THE INDEX
--CREATED BY ARCGIS
[S1169_idx] 
ON 
--AND THIS IS THE NAME OF TABLE
--SUPPLIED BY THE USER IN ARCGIS
--AT CREATION TIME
[dbo].[TEST_GEOM]
(
  [SHAPE]
)USING  GEOMETRY_GRID 
WITH (BOUNDING_BOX =(227166.13, 3925740.74, 314851.6915, 3968047.64), 
GRIDS =(
LEVEL_1 = MEDIUM,
LEVEL_2 = MEDIUM,
LEVEL_3 = MEDIUM,
LEVEL_4 = MEDIUM), 
CELLS_PER_OBJECT = 16, PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, SORT_IN_TEMPDB = OFF,
--DROP_EXISTING = ON ALLOWS US TO DO THIS ALL
--DAY WITHOUT MESSING WITH DROPPING THE IDX 
DROP_EXISTING = ON, 
ONLINE = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON, FILLFACTOR = 100) ON [PRIMARY]
GO
DECLARE @geom geometry
--THE POLYGON IS AN AREA SMALLER THAN THE BOUNDING BOX
SET @geom = GEOMETRY::STGeomFromText('POLYGON ((247804.201 3943957.896, 29932.568 3943963.210, 247671.344 3942876.441, 247684.630 3943652.325,247804.201 3943957.896))', 26917)jjj
exec sp_help_spatial_geometry_index 'TEST_GEOM', 'S1169_idx', 1, @geom

There's a WHOLE lot of output. The first thing my eye spies is

and

I'm no expert on SQL or spatial indexes, but I don't think those numbers look good. No wonder my draw times are slow, spatial queries are sluggish, and in general, the database is performing horribly. Every single ArcGIS-created table has the same index parameters!

Just for giggle, lets change the cells per object by altering just this in the above code snippet:

CELLS_PER_OBJECT = 4096

And look at the increase in efficiencies!

Again, I'm claiming to be no expert at SQL Spatial Index Tuning, but I think I'm on to something here. As it turns out, my test data consists of "Very Complex Line Strings", which, if you're not living under the GIS rock, that's what all of your data is. Coincidentally, using value of 8192 for Cells Per Object in this scenario is a good starting point.

But there's a lot (lots and lots) of values between 16 and 8092, and then there's all the permutations of Low, Medium, and High that could be tested to determine the spatial index that is likely to give you the best performance most of the time. What if there was a way to automatically test spatial index settings and magically determine which parameters best fit your data scenario?

Enter the SDE Hacker....

I stumbled across this geospatial - Selecting a good SQL Server 2008 spatial index with large polygons - Stack Overflow  post where someone had posted a SQL stored procedure that will loop through Cell Sizes and Grid Levels to report Spatial Query results for every permutation (user supplied). Being the tinker that I am, I of course broke the code trying to add a bunch of output. Finally got it working. Here it is:

--ORIGINAL SQL CODE FROM
--http://stackoverflow.com/users/2250424/greengeo
--MODIFIFIED TO INCLUDE SPATIAL INDEX HELP OUTPUT
--IN TABLE
  --!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  --LOOK FOR THE EXCLAMATION POINTS IN THIS CODE
  --THERE ARE SEVERAL USER-SUPPLIED VARIABLES THAT MUST
  --BE INPUT FOR THIS SP TO WORK
  --!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!


USE DATBASE
GO



CREATE PROCEDURE dbo.sp_tune_spatial_index
(
  @tabnm VARCHAR(MAX), -- This parameter stores the name of the spatial table for which you are tuning the index
  @idxnm VARCHAR(MAX), -- This parameter stores the name of the spatial index of the named table
  @min_cells_per_obj INT, -- Minimum Cells Per Object to test on. Suggested to start at 2.
  @max_cells_per_obj INT, -- Maximum Cells Per Object to test on.

  /* The test requires two geometry instances to use in test query 1 and 2.
  The first one should cover the area of default extent. The second should
  cover an area roughly the size of the area shown when zoomed in, panning
  around. It is required that the variable store a string that will create
  the geometry instance since this will be done within the procedure and 
  cannot be a variable of type: GEOMETRY. The SRID of these instances must
  match that of the table you are testing. */
  @testgeom1 VARCHAR(MAX), -- This parameter stores the first geometry instance creation string that will be used in the test
  @testgeom2 VARCHAR(MAX) -- This parameter stores the second geometry instance creation string that will be used in the test



)

AS

SET NOCOUNT ON;

/* Prior to running this procedure, two tables are required. These tables are 
  created here to prepare for running the procedure. */

PRINT 'Checking for required tables...'
IF EXISTS(SELECT 1 FROM sysobjects WHERE name IN ('cell_opt_perm', 'spat_idx_test_result'))
  BEGIN
  PRINT '... The "cell_opt_perm" and "spat_idx_test_result" tables exist.'
  END
ELSE
BEGIN
  PRINT '... Creating "cell_opt_perm" and "spat_idx_test_result" tables.'
  CREATE TABLE cell_opt_perm(
  [perm_id] [smallint] NOT NULL,
  [permutation] [nvarchar](4) NOT NULL,
  [level1] [nvarchar](6) NOT NULL,
  [level2] [nvarchar](6) NOT NULL,
  [level3] [nvarchar](6) NOT NULL,
  [level4] [nvarchar](6) NOT NULL
  )

  INSERT INTO cell_opt_perm ([perm_id], [permutation], [level1], [level2], [level3], [level4])
  VALUES (1,'LLLL','LOW','LOW','LOW','LOW'),
  (2,'LLLM','LOW','LOW','LOW','MEDIUM'),
  (3,'LLLH','LOW','LOW','LOW','HIGH'),
  (4,'LLML','LOW','LOW','MEDIUM','LOW'),
  (5,'LLMM','LOW','LOW','MEDIUM','MEDIUM'),
  (6,'LLMH','LOW','LOW','MEDIUM','HIGH'),
  (7,'LLHL','LOW','LOW','HIGH','LOW'),
  (8,'LLHM','LOW','LOW','HIGH','MEDIUM'),
  (9,'LLHH','LOW','LOW','HIGH','HIGH'),
  (10,'LMLL','LOW','MEDIUM','LOW','LOW'),
  (11,'LMLM','LOW','MEDIUM','LOW','MEDIUM'),
  (12,'LMLH','LOW','MEDIUM','LOW','HIGH'),
  (13,'LMML','LOW','MEDIUM','MEDIUM','LOW'),
  (14,'LMMM','LOW','MEDIUM','MEDIUM','MEDIUM'),
  (15,'LMMH','LOW','MEDIUM','MEDIUM','HIGH'),
  (16,'LMHL','LOW','MEDIUM','HIGH','LOW'),
  (17,'LMHM','LOW','MEDIUM','HIGH','MEDIUM'),
  (18,'LMHH','LOW','MEDIUM','HIGH','HIGH'),
  (19,'LHLL','LOW','HIGH','LOW','LOW'),
  (20,'LHLM','LOW','HIGH','LOW','MEDIUM'),
  (21,'LHLH','LOW','HIGH','LOW','HIGH'),
  (22,'LHML','LOW','HIGH','MEDIUM','LOW'),
  (23,'LHMM','LOW','HIGH','MEDIUM','MEDIUM'),
  (24,'LHMH','LOW','HIGH','MEDIUM','HIGH'),
  (25,'LHHL','LOW','HIGH','HIGH','LOW'),
  (26,'LHHM','LOW','HIGH','HIGH','MEDIUM'),
  (27,'LHHH','LOW','HIGH','HIGH','HIGH'),
  (28,'MLLL','MEDIUM','LOW','LOW','LOW'),
  (29,'MLLM','MEDIUM','LOW','LOW','MEDIUM'),
  (30,'MLLH','MEDIUM','LOW','LOW','HIGH'),
  (31,'MLML','MEDIUM','LOW','MEDIUM','LOW'),
  (32,'MLMM','MEDIUM','LOW','MEDIUM','MEDIUM'),
  (33,'MLMH','MEDIUM','LOW','MEDIUM','HIGH'),
  (34,'MLHL','MEDIUM','LOW','HIGH','LOW'),
  (35,'MLHM','MEDIUM','LOW','HIGH','MEDIUM'),
  (36,'MLHH','MEDIUM','LOW','HIGH','HIGH'),
  (37,'MMLL','MEDIUM','MEDIUM','LOW','LOW'),
  (38,'MMLM','MEDIUM','MEDIUM','LOW','MEDIUM'),
  (39,'MMLH','MEDIUM','MEDIUM','LOW','HIGH'),
  (40,'MMML','MEDIUM','MEDIUM','MEDIUM','LOW'),
  (41,'MMMM','MEDIUM','MEDIUM','MEDIUM','MEDIUM'),
  (42,'MMMH','MEDIUM','MEDIUM','MEDIUM','HIGH'),
  (43,'MMHL','MEDIUM','MEDIUM','HIGH','LOW'),
  (44,'MMHM','MEDIUM','MEDIUM','HIGH','MEDIUM'),
  (45,'MMHH','MEDIUM','MEDIUM','HIGH','HIGH'),
  (46,'MHLL','MEDIUM','HIGH','LOW','LOW'),
  (47,'MHLM','MEDIUM','HIGH','LOW','MEDIUM'),
  (48,'MHLH','MEDIUM','HIGH','LOW','HIGH'),
  (49,'MHML','MEDIUM','HIGH','MEDIUM','LOW'),
  (50,'MHMM','MEDIUM','HIGH','MEDIUM','MEDIUM'),
  (51,'MHMH','MEDIUM','HIGH','MEDIUM','HIGH'),
  (52,'MHHL','MEDIUM','HIGH','HIGH','LOW'),
  (53,'MHHM','MEDIUM','HIGH','HIGH','MEDIUM'),
  (54,'MHHH','MEDIUM','HIGH','HIGH','HIGH'),
  (55,'HLLL','HIGH','LOW','LOW','LOW'),
  (56,'HLLM','HIGH','LOW','LOW','MEDIUM'),
  (57,'HLLH','HIGH','LOW','LOW','HIGH'),
  (58,'HLML','HIGH','LOW','MEDIUM','LOW'),
  (59,'HLMM','HIGH','LOW','MEDIUM','MEDIUM'),
  (60,'HLMH','HIGH','LOW','MEDIUM','HIGH'),
  (61,'HLHL','HIGH','LOW','HIGH','LOW'),
  (62,'HLHM','HIGH','LOW','HIGH','MEDIUM'),
  (63,'HLHH','HIGH','LOW','HIGH','HIGH'),
  (64,'HMLL','HIGH','MEDIUM','LOW','LOW'),
  (65,'HMLM','HIGH','MEDIUM','LOW','MEDIUM'),
  (66,'HMLH','HIGH','MEDIUM','LOW','HIGH'),
  (67,'HMML','HIGH','MEDIUM','MEDIUM','LOW'),
  (68,'HMMM','HIGH','MEDIUM','MEDIUM','MEDIUM'),
  (69,'HMMH','HIGH','MEDIUM','MEDIUM','HIGH'),
  (70,'HMHL','HIGH','MEDIUM','HIGH','LOW'),
  (71,'HMHM','HIGH','MEDIUM','HIGH','MEDIUM'),
  (72,'HMHH','HIGH','MEDIUM','HIGH','HIGH'),
  (73,'HHLL','HIGH','HIGH','LOW','LOW'),
  (74,'HHLM','HIGH','HIGH','LOW','MEDIUM'),
  (75,'HHLH','HIGH','HIGH','LOW','HIGH'),
  (76,'HHML','HIGH','HIGH','MEDIUM','LOW'),
  (77,'HHMM','HIGH','HIGH','MEDIUM','MEDIUM'),
  (78,'HHMH','HIGH','HIGH','MEDIUM','HIGH'),
  (79,'HHHL','HIGH','HIGH','HIGH','LOW'),
  (80,'HHHM','HIGH','HIGH','HIGH','MEDIUM'),
  (81,'HHHH','HIGH','HIGH','HIGH','HIGH')

  CREATE TABLE spat_idx_test_result(
  [perm_id] [int] NOT NULL,
  [num_cells] [int] NOT NULL,
  [permut] [nvarchar](4) NOT NULL,
  [g1t1] [bigint] NULL,
  [g1t2] [bigint] NULL,
  [g1t3] [bigint] NULL,
  [g1t4] [bigint] NULL,
  [g2t1] [bigint] NULL,
  [g2t2] [bigint] NULL,
  [g2t3] [bigint] NULL,
  [g2t4] [bigint] NULL,
  [PF_EFF][float] NULL,
  [IF_EFF][float] NULL,
  [GRIDL1] [int] NULL,
  [GRIDL2] [int] NULL,
  [GRIDL3] [int] NULL,
  [GRIDL4] [int] NULL,
  [TPIR] [bigint] NULL,
  [TPIP] [bigint] NULL,
  [ANOIRPBR] [bigint] NULL,
  [TNOOCILFQ] [bigint] NULL,
  [TNOOCIL3FQ] [bigint] NULL,
  [TNOOCIL4FQ] [bigint] NULL,
  [TNOOCIL0II] [bigint] NULL,
  [TNOOCIL4II] [bigint] NULL,
  [TNOIOIL3FQ] [bigint] NULL,
  [TNOIOIL4FQ] [bigint] NULL,
     [ITTCNTLGP] [float] NULL,
  [INTTTCNTLGP] [float] NULL,
  [BTTCNTLGP] [float] NULL,
  [ACPONTLGP] [float] NULL,
  [AOPG] [float] NULL,
  [NORSBPF] [bigint] NULL,
  [NORSBIF] [bigint] NULL,
  [NOTSFIC] [bigint] NULL,
  [NORO] [bigint] NULL,
  [PORNBPF] [float] NULL,
  [POPFRSBIF] [float] NULL


  )

  INSERT INTO dbo.spat_idx_test_result
  VALUES (0,16,0,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL)
END


/* delete all rows from "spat_idx_test_result" table. This makes it ready to stuff in new results.
  !!!WARNING!!! if your test was interupted, the table will be cleared out and the test will
  begin from the beginning. You could try to modify this to start where you left off but
  I didn't have time and this worked well enough for me. */
DELETE FROM spat_idx_test_result
  WHERE perm_id != 0

/* set up counters */
DECLARE @a1 INT
DECLARE @a2 INT
DECLARE @a3 INT
DECLARE @a4 INT

/* set up variables to hold high/medium/low values and permutation to use in rebuilding
   the spatial index and recording stats */
DECLARE @lev1 VARCHAR(6)
DECLARE @lev2 VARCHAR(6)
DECLARE @lev3 VARCHAR(6)
DECLARE @lev4 VARCHAR(6)
DECLARE @permut VARCHAR(6)
DECLARE @num_cell VARCHAR(4)
DECLARE @time_str VARCHAR(20)
DECLARE @perm_id VARCHAR(20)
DECLARE @x xml
DECLARE @pf_eff FLOAT



/* create variables to hold timestamps for beginning and ending of test queries */
DECLARE @start_t DATETIME
DECLARE @end_t DATETIME
DECLARE @elapse_t INT



/* begin looping through cell option permutations */
SET @a1 = @min_cells_per_obj
WHILE @a1 <= @max_cells_per_obj
  BEGIN
  SET @a2 = 1
  PRINT 'Started Testing for ' +CAST(@a1 AS VARCHAR(10)) +' cells per object'
  WHILE @a2 < 82
  BEGIN
  SELECT @lev1 = level1, @lev2 = level2, @lev3 = level3, @lev4 = level4 FROM cell_opt_perm WHERE perm_id = @a2
  SET @permut = '''' +(SELECT permutation FROM cell_opt_perm WHERE perm_id = @a2) +''''


  EXEC
  ('
  CREATE SPATIAL INDEX ' +@idxnm +' ON ' +@tabnm +' 
  (
  [SHAPE]
  )
  USING  GEOMETRY_GRID 
  WITH
  (

  --!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  --MAKE SURE YOU EDIT THE BOUNDING BOX TO BE EXACTLY EQUAL TO THE BOUNDING
  --BOX OF YOUR SPATIAL TABLE 
  --!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

  BOUNDING_BOX =(227166.13, 3925740.74, 314851.6915, 3968047.64),
  GRIDS =(LEVEL_1 = ' +@lev1 +' ,LEVEL_2 = ' +@lev2 +' ,LEVEL_3 = ' +@lev3 +' ,LEVEL_4 = ' +@lev4 +' ), 
  CELLS_PER_OBJECT = ' +@a1 +' ,
  PAD_INDEX  = OFF,
  SORT_IN_TEMPDB = OFF,
  DROP_EXISTING = ON,
  ALLOW_ROW_LOCKS  = ON,
  ALLOW_PAGE_LOCKS  = ON,
  FILLFACTOR = 100
  )
  ON [PRIMARY]
  '
  )


  PRINT 'Re-built index to ' +@permut
  SET @a3 = 1
  SET @a4 = 1
  WHILE @a3 < 5
  BEGIN
  SET @start_t = GETDATE()
  EXEC
  (
  'CREATE TABLE #tmp_tab (shp GEOMETRY)
  DECLARE @g1 GEOMETRY
  SET @g1 = ' +@testgeom1 +'
  INSERT #tmp_tab (shp)
  SELECT
  r.Shape AS shp
  FROM
  ' +@tabnm +' r
  WHERE
  r.SHAPE.STIntersects(@g1) = 1
  DROP TABLE #tmp_tab'
  )
  SET @end_t = GETDATE()
  SET @elapse_t = (SELECT DATEDIFF(MS, @start_t, @end_t))
  SET @num_cell = CAST(@a1 AS VARCHAR(6))
  SET @time_str = CAST(@elapse_t AS VARCHAR(20))
  IF @a3 = 1
  BEGIN
  IF (SELECT TOP 1 perm_id FROM spat_idx_test_result) IS NULL
  BEGIN
  SET @perm_id = 1


  END
  ELSE
  BEGIN
  SET @perm_id = CAST((SELECT MAX(perm_id+1) FROM spat_idx_test_result) AS VARCHAR(20))
  END


  EXEC


  (
  '


  INSERT INTO spat_idx_test_result (perm_id, num_cells, permut, g1t' +@a3 +')
  VALUES (' +@perm_id +', ' +@num_cell +', ' +@permut +', ' +@time_str +')'
  )
  END
  ELSE


  EXEC
  (
  '

  UPDATE spat_idx_test_result
  SET
  num_cells = ' +@num_cell +',
  permut = ' +@permut +',
  g1t' +@a3 +' = ' +@time_str +'
  WHERE perm_id = ' +@perm_id 


  )
  SET @a3 = @a3 + 1
  END
  WHILE @a4 < 5
  BEGIN

  SET @start_t = GETDATE()
  EXEC
  (
  'CREATE TABLE #tmp_tab (shp GEOMETRY) 
  DECLARE @g2 GEOMETRY
  SET @g2 = ' +@testgeom2 +'
  INSERT #tmp_tab (shp)
  SELECT
  r.Shape AS shp
  FROM
  ' +@tabnm +' r
  WHERE
  r.SHAPE.STIntersects(@g2) = 1
  DROP TABLE #tmp_tab'
  )
  SET @end_t = GETDATE()
  SET @elapse_t = (SELECT DATEDIFF(MS, @start_t, @end_t))
  SET @num_cell = CAST(@a1 AS VARCHAR(6))
  SET @time_str = CAST(@elapse_t AS VARCHAR(20))
  EXEC
  (
  '
  DECLARE @geom geometry
  DECLARE @x xml
  DECLARE @PFVALUE float
  DECLARE @IFVALUE float
  DECLARE @GRIDL1VALUE int
  DECLARE @GRIDL2VALUE int
  DECLARE @GRIDL3VALUE int
  DECLARE @GRIDL4VALUE int
  DECLARE @TPIRVALUE bigint
  DECLARE @TPIPVALUE bigint
  DECLARE @ANOIRPBRVALUE bigint
  DECLARE @TNOOCILFQVALUE bigint
  DECLARE @TNOOCIL0IIVALUE bigint
  DECLARE @TNOOCIL4IIVALUE bigint
  DECLARE @TNOOCIL3FQVALUE bigint
  DECLARE @TNOOCIL4FQVALUE bigint
  DECLARE @TNOIOIL3FQVALUE bigint
  DECLARE @TNOIOIL4FQVALUE bigint
  DECLARE @ITTCNTLGPVALUE float
  DECLARE @INTTTCNTLGPVALUE float
  DECLARE @BTTCNTLGPVALUE float
  DECLARE @ACPONTLGPVALUE float
  DECLARE @AOPGVALUE float
  DECLARE @NORSBPFVALUE bigint
  DECLARE @NORSBIFVALUE bigint
  DECLARE @NOTSFICVALUE bigint
  DECLARE @NOROVALUE bigint
  DECLARE @PORNBPFVALUE float
  DECLARE @POPFRSBIFVALUE float
  --!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  --MAKE SURE YOU EDIT GEOMETRY VARIABLE BELOW TO REPRESENT A POLYGON
  --THAT IS WITHIN YOUR BOUNDING BOX
  --!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  SET @geom = GEOMETRY::STGeomFromText(''POLYGON ((247804.201 3943957.896, 29932.568 3943963.210, 247671.344 3942876.441, 247684.630 3943652.325,247804.201 3943957.896))'', 26917)
  --!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  --MAKE SURE YOU SPECIFY THE NAME OF YOUR SPATIAL TABLE
  --AND THE NAME OF THE SPATIAL INDEX
  -- IN THE sp_help_spatial_geometry_index_xml VARIABLES
  --!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
  exec sp_help_spatial_geometry_index_xml TEST_GEOM , S1169_idx , 1, @geom, @x output
  SET @PFVALUE =  @x.value(''(/Primary_Filter_Efficiency/text())[1]'', ''float'')
  SET @IFVALUE =  @x.value(''(/Internal_Filter_Efficiency/text())[1]'', ''float'')
  SET @GRIDL1VALUE =  @x.value(''(/Grid_Size_Level_1/text())[1]'', ''int'')
  SET @GRIDL2VALUE =  @x.value(''(/Grid_Size_Level_2/text())[1]'', ''int'')
  SET @GRIDL3VALUE =  @x.value(''(/Grid_Size_Level_3/text())[1]'', ''int'')
  SET @GRIDL4VALUE =  @x.value(''(/Grid_Size_Level_4/text())[1]'', ''int'')
  SET @TPIRVALUE =  @x.value(''(/Total_Primary_Index_Rows/text())[1]'', ''bigint'')
  SET @TPIPVALUE =  @x.value(''(/Total_Primary_Index_Pages/text())[1]'', ''bigint'')
  SET @ANOIRPBRVALUE =  @x.value(''(/Average_Number_Of_Index_Rows_Per_Base_Row/text())[1]'', ''bigint'')
  SET @TNOOCILFQVALUE =  @x.value(''(/Total_Number_Of_ObjectCells_In_Level0_For_QuerySample/text())[1]'', ''bigint'')
  SET @TNOOCIL0IIVALUE =  @x.value(''(/Total_Number_Of_ObjectCells_In_Level0_In_Index/text())[1]'', ''bigint'')
  SET @TNOOCIL4IIVALUE =  @x.value(''(/Total_Number_Of_ObjectCells_In_Level4_In_Index/text())[1]'', ''bigint'')
  SET @TNOOCIL3FQVALUE =  @x.value(''(/Total_Number_Of_ObjectCells_In_Level3_For_QuerySample/text())[1]'', ''bigint'')
  SET @TNOOCIL4FQVALUE =  @x.value(''(/Total_Number_Of_ObjectCells_In_Level4_For_QuerySample/text())[1]'', ''bigint'')
  SET @TNOIOIL3FQVALUE =  @x.value(''(/Total_Number_Of_Interior_ObjectCells_In_Level3_For_QuerySample/text())[1]'', ''bigint'')
  SET @TNOIOIL4FQVALUE =  @x.value(''(/Total_Number_Of_Interior_ObjectCells_In_Level4_For_QuerySample/text())[1]'', ''bigint'')
  SET @ITTCNTLGPVALUE =  @x.value(''(/Interior_To_Total_Cells_Normalized_To_Leaf_Grid_Percentage/text())[1]'', ''float'')
  SET @INTTTCNTLGPVALUE =  @x.value(''(/Intersecting_To_Total_Cells_Normalized_To_Leaf_Grid_Percentage/text())[1]'', ''float'')
  SET @BTTCNTLGPVALUE =  @x.value(''(/Border_To_Total_Cells_Normalized_To_Leaf_Grid_Percentage/text())[1]'', ''float'')
  SET @ACPONTLGPVALUE =  @x.value(''(/Average_Cells_Per_Object_Normalized_To_Leaf_Grid/text())[1]'', ''float'')
  SET @AOPGVALUE =  @x.value(''(/Average_Objects_PerLeaf_GridCell/text())[1]'', ''float'')
  SET @NORSBPFVALUE =  @x.value(''(/Number_Of_Rows_Selected_By_Primary_Filter/text())[1]'', ''bigint'')
  SET @NORSBIFVALUE =  @x.value(''(/Number_Of_Rows_Selected_By_Internal_Filter/text())[1]'', ''bigint'')
  SET @NOTSFICVALUE =  @x.value(''(/Number_Of_Times_Secondary_Filter_Is_Called/text())[1]'', ''bigint'')
  SET @NOROVALUE =  @x.value(''(/Number_Of_Rows_Output/text())[1]'', ''bigint'')
  SET @PORNBPFVALUE =  @x.value(''(/Percentage_Of_Rows_NotSelected_By_Primary_Filter/text())[1]'', ''float'')
  SET @POPFRSBIFVALUE =  @x.value(''(/Percentage_Of_Primary_Filter_Rows_Selected_By_Internal_Filter/text())[1]'', ''float'')


  UPDATE spat_idx_test_result
  SET
  num_cells = ' +@num_cell +',
  permut = ' +@permut +',
  g2t' +@a4 +' = ' +@time_str +',
  PF_EFF = @PFVALUE,
  IF_EFF = @IFVALUE,
  GRIDL1 = @GRIDL1VALUE,
  GRIDL2 = @GRIDL2VALUE,
  GRIDL3 = @GRIDL3VALUE,
  GRIDL4 = @GRIDL4VALUE,
  TPIR = @TPIRVALUE,
  TPIP = @TPIPVALUE,
  ANOIRPBR = @ANOIRPBRVALUE,
  TNOOCILFQ = @TNOOCILFQVALUE,
  TNOOCIL0II = @TNOOCIL0IIVALUE,
  TNOOCIL4II = @TNOOCIL4IIVALUE,
  TNOOCIL3FQ =  @TNOOCIL3FQVALUE,
  TNOOCIL4FQ =  @TNOOCIL4FQVALUE,
  TNOIOIL3FQ = @TNOIOIL3FQVALUE,
  TNOIOIL4FQ = @TNOIOIL4FQVALUE,
  ITTCNTLGP = @ITTCNTLGPVALUE,
  INTTTCNTLGP = @INTTTCNTLGPVALUE,
  BTTCNTLGP = @BTTCNTLGPVALUE,
  ACPONTLGP = @ACPONTLGPVALUE,
  AOPG = @AOPGVALUE,
  NORSBPF = @NORSBPFVALUE,
  NORSBIF = @NORSBIFVALUE,
  NOTSFIC = @NOTSFICVALUE,
  NORO = @NOROVALUE,
  PORNBPF = @PORNBPFVALUE,
  POPFRSBIF = @POPFRSBIFVALUE
  WHERE perm_id = ' +@perm_id
  )
  SET @a4 = @a4 + 1
  END
  SET @a2 = @a2 + 1
  END
  SET @a1 = @a1 + 1

  END
PRINT 'Testing of ' +@tabnm +' spatial index: ' +@idxnm +' is complete!'
GO

The hacking that I did involved using the xml output of the index help stored procedure to write some more descriptive results to the output in addition to query time.

The stored procedure is executed like so:

DECLARE @BOUNDING VARCHAR(MAX) 
SET @BOUNDING = 'GEOMETRY::STGeomFromText(''POLYGON ((226805.072 3975572.527, 318101.215 3975985.165, 317894.896 3922239.073, 225360.839  3926571.772 , 226805.072 3975572.527))'', 0)'

DECLARE @QUERY VARCHAR(MAX) 
SET @QUERY = 'GEOMETRY::STGeomFromText(''POLYGON ((247804.201 3943957.896, 29932.568 3943963.210, 247671.344 3942876.441, 247684.630 3943652.325, 247804.201 3943957.896))'', 26917)'

EXEC sp_tune_spatial_index 'TEST_GEOM', 'S1169_idx', 4096, 4096, @BOUNDING, @QUERY 
GO

In this example I'm only testing one Cell Size: 4096, but you could use any value range such as 8, 16.

Results can be nicely reviewed with the following:

SELECT  
    perm_id as 'Permutation #'
      ,num_cells 'Cells Per Object'
      ,permut as 'Grids'
      ,g1t1 as 'ms to query entire geometry (Level 1)'
      ,g1t2 as 'ms to query entire geometry (Level 2)'
      ,g1t3 as 'ms to query entire geometry (Level 3)'
      ,g1t4 as 'ms to query entire geometry (Level 4)'
      ,g2t1 as 'ms to execute spatial query (Level 1)'
      ,g2t2 as 'ms to execute spatial query (Level 2)'
      ,g2t3 as 'ms to execute spatial query (Level 3)'
      ,g2t4 as 'ms to execute spatial query (Level 4)'
      ,PF_EFF as 'Primary Filter Efficiency'
      ,IF_EFF as 'Internal Filter Efficiency'
      ,GRIDL1 as 'Grid Size Level 1'
      ,GRIDL2 as 'Grid Size Level 2'
      ,GRIDL3 as 'Grid Size Level 3'
      ,GRIDL4 as 'Grid Size Level 4'
      ,TPIR as 'Total Primary Index Rows'
      ,TPIP as 'Total Primary Index Pages'
      ,ANOIRPBR as 'Average Number of Index Rows Per Base Row'
      ,TNOOCILFQ as 'Total Number of Object Cells in Level 0 For Query Sample'
   ,TNOOCIL3FQ as 'Total Number of Object Cells in Level 3 For Query Sample'
   ,TNOOCIL4FQ as 'Total Number of Object Cells in Level 4 For Query Sample'
      ,TNOOCIL0II as 'Total Number of Object Cells In Level 0 In Index'
      ,TNOOCIL4II as 'Total Number of Object Cells In Level 4 In Index'
   ,TNOIOIL3FQ as 'Total Number Of Interior ObjectCells In Level 3 For QuerySample'
   ,TNOIOIL4FQ as 'Total Number Of Interior ObjectCells In Level 4 For QuerySample'
   ,ITTCNTLGP as 'Interior To Total Cells Normalized To Leaf Grid Percentage'
   ,INTTTCNTLGP as 'Intersecting To Total Cells Normalized To Leaf Grid Percentage'
   ,BTTCNTLGP as 'Border To Total Cells Normalized To Leaf Grid Percentage'
   ,ACPONTLGP as 'Average Cells Per Object Normalized To Leaf Grid'
   ,AOPG as 'Average Objects PerLeaf GridCell'
   ,NORSBPF as 'Number Of Rows Selected By Primary Filter'
   ,NORSBIF as 'Number Of Rows Selected By Internal Filter'
   ,NOTSFIC as 'Number Of Times Secondary Filter Is Called'
   ,NORO as 'Number Of Rows Output'
   ,PORNBPF as 'Percentage Of Rows NotSelected By Primary Filter'
   ,POPFRSBIF as 'Percentage Of Primary Filter Rows Selected By Internal Filter'
FROM spat_idx_test_result

ORDER BY PF_EFF

This post doesn't even scratch the surface of Spatial Index Tuning. For example, if you're using SQL 2012, you could set the index to Auto Grid, which gives you 8 Cell Levels and supposedly is a good marriage for ArcGIS. You also want to look at other performance factors such as Internal Filter Efficiency. There's a lot of Witch Magic and Snake Oil Medicine to Spatial Index Tuning. Hope this blog post gets you started in the right direction!

In Real-time Geodatabase Replication? Part 1 you were introduced to SQL Real-Time Replication. One requirement of SQL Merge Replication is that the table being replicated must have a rowguidcol. This is easy with a cursor:

DECLARE 
@sql VARCHAR(500), 
@tableName VARCHAR(128)
DECLARE gandgCursor CURSOR 
FOR 
SELECT 
table_name
FROM 
SDE_column_registry
WHERE 
table_name like 'CR%' 
ORDER BY 
table_name ASC 
OPEN gandgCursor 
FETCH NEXT FROM gandgCursor  
INTO @tableName 
WHILE ( @@FETCH_STATUS = 0 ) 
BEGIN 
SET @sql = ' 
ALTER TABLE [dbo].['+ @tableName + '] 
ALTER COLUMN GlobalID ADD ROWGUIDCOL ;
' 
 PRINT 'Executing Statement - '+ @sql 
 EXECUTE ( @sql ) 
 FETCH NEXT FROM gandgCursor  
 INTO  @tableName 
 END 
 CLOSE gandgCursor 
 DEALLOCATE gandgCursor

Messin' up the scenery, breakin' my mind. Do this, don't do that, can't you read the LINE? In Hey Neighbor? What's Your Value? , Rasters, Pixels and Code Oh My!, and Using SQL to Auto-Populate XY Values I had introduced some SQL trickery to automate some data entry. Probably should have mentioned those methods mostly just work on points....

For our GIS data standard, we also want to collect the same administrative values and automation for line types such as roads, trails, and other linear features. Problem is, there's a lot of vertices (nodes, coordinate pairs) on a line, which one to use for the SQL trickery?

There are some existing SQL Geography methods for manipulating well-known-text from a Geography line instance, but none of them give me a midpoint value, and they certainly don't lend themselves well to being part of a trigger.

In order to easily get a line midpoint value from Geography, you'll need the SQL Server Spatial Tools - Home  installed, which is as easy as registering a DLL through SQL on your database instance- works on 2008 and 2012.

A line midpoint can be queried as

SELECT dbo.LocateAlongGeog(SHAPE, SHAPE.STLength()/2).Lat FROM [dbo].[SOMELINE]
SELECT dbo.LocateAlongGeog(SHAPE, SHAPE.STLength()/2).Long FROM [dbo].[SOMELINE]

Wrapping that up in a trigger to populate the LAT, LON, and Elevation fields we get:

CREATE TRIGGER    [dbo].[SOMELINE_GEO]
ON    [dbo].[SOMELINE]
AFTER INSERT, UPDATE NOT FOR REPLICATION
AS   
BEGIN    
SET NOCOUNT ON;  
 UPDATE p SET  
LON = CASE WHEN p.SHAPE IS NULL THEN p.STStartPoint.LON ELSE p.SHAPE.Long END,  
LAT = CASE WHEN p.SHAPE IS NULL THEN p.STStartPoint.LAT ELSE p.SHAPE.Lat END,
ELEVATION = (SELECT
pdata.getValueByLoc(1,p.SHAPE.Long,p.SHAPE.Lat)  FROM [dbo].[DEM10MP])
  FROM  
    dbo.SOMELINE AS p 
  INNER JOIN  
    inserted AS i 
    ON i.GlobalID = p.GlobalID 
END
GO

Polygons don't require the "special function" and simply can be queried for midpoint values with:

SELECT SHAPE.EnvelopeCenter().Lat
FROM [SOMEPOLY
SELECT SHAPE.EnvelopeCenter().Long
FROM [SOMEPOLY]

It's very easy as a GIS administrator to add lots and lots of fields to a feature class and even easier to naively think that folks are going to populate or edit them! One common theme for me is date fields. We have edit date, create date, FGDC edit and create date, year...blah blah blah.

I wish there was just one, or even no, onus on editors to have to think about dates and the database just...handled it.

This can be so. With a database trigger. Let's start with FGDC dates. If you're implementing some form of feature-level metadata, or some data-mining that tags another metadata element with FGDC source, create or edit date yet have editor tracking enabled, there's no reason why you should have to also populate the FGDC[whatever]DATE column as well.

CREATE TRIGGER [dbo].[SOMETABLE_DATE]
ON [dbo].[SOMETABLE]
AFTER INSERT, UPDATE NOT FOR REPLICATION
AS BEGIN
SET NOCOUNT ON; 
UPDATE [dbo].[SOMETABLE]
SET 
SRCDATEFGDC = (convert(varchar(8), SOURCEDATE, 112)), 
CREATEDATEFGDC = (convert(varchar(8), CREATEDATE, 112)), 
EDITDATEFGDC = (convert(varchar(8), EDITDATE, 112))
END
GO

Here we're taking the EDITDATE and CREATE date values, which are sql

datetime(2)

and converting them to a string format as yyyymmdd. The 112 is what controls the output format. See CAST and CONVERT (Transact-SQL) for a full list of date conversion formats.

In addition, the user is also selecting a source date in this case which may be different than create date and that is converted as well.

If you have a year column you can also

YEAR = CASE WHEN SOURCEDATE IS NULL THEN NULL ELSE YEAR(SOURCEDATE) END

with null value handling thrown in.

If you have editor tracking disabled for some reason (often causes issues with Collector for ArcGIS) you could through a default constraint on your date column(s) as

getdate()

.

This is a personal blog and does not recommend, endorse, or support the methods described above. Alteration of data using SQL outside of the ESRI software stack, of course, is not supported and should not be applied to a production database without a thorough understanding and disaster recovery plan.

Do you have someone in your organization that YELLS WITH THEIR KEYBOARD? Everything is upper case? What about the rogue all-lower-case folks? Or ever worse, the First-word-is-proper-case-all-other-words-are-lower-case ninjas.

I have a personal pet-peeve (OCD). Feature names in GIS should be Proper Case. Happy Valley Road. Not Happy valley road, not Happy valley Road, and definitely not HAPPY VALLEY ROAD.

I'd like to enforce Proper Case naming of those feature regardless of how the text is cased by the editor. This can be accomplished with a simple Function and Trigger in SQL.

First create the following Function:

create function [dbo].[ProperCase](@Text as varchar(8000))
returns varchar(8000)
as
begin
   declare @Reset bit;
   declare @Ret varchar(8000);
   declare @i int;
   declare @c char(1);


   select @Reset = 1, @i=1, @Ret = '';
   
   while (@i <= len(@Text))
    select @c= substring(@Text,@i,1),
               @Ret = @Ret + case when @Reset=1 then UPPER(@c) else LOWER(@c) end,
               @Reset = case when @c like '[a-zA-Z]' then 0 else 1 end,
               @i = @i +1
   return @Ret
end
GO

Then this Triggger:

CREATE TRIGGER [dbo].[NAME_UPDATE]
ON [dbo].[SOME_TABLE]
AFTER INSERT NOT FOR REPLICATION
AS BEGIN
SET NOCOUNT ON; 
UPDATE SOME_TABLE
SET
NAME = dbo.Propercase(NAME)
END
GO

Note here that this only fires after an insert, not an update. There could be legitimate reason why the YELLERS want something other than proper case, e.g. "ND Happy Valley Road" (ND for "North District). This allows them (or me after they YELL at me) to update that one feature without the trigger proper-casing my edit.

This is a personal blog and does not recommend, endorse, or support the methods described above. Alteration of data using SQL outside of the ESRI software stack, of course, is not supported and should not be applied to a production database without a thorough understanding and disaster recovery plan.

I have tried, and failed, many times to get geodatabase replication using ESRI tools to work, work reliably, or work at all. Perhaps it's all the un-authorized SQL tinkering I do. Even when I do get it working, I'm not a big fan of having to fire up Arc Toolbox and push buttons to make replication happen. Not a big fan of writing Python code, either.

With SQL server, there IS a way to replicate an entire database, or parts of of a database. The part of a database is handled with Merge Replication, in which you define what tables to replicate, and what criteria under which data is replicated. I like that option! Consider this environment:

I have a water quality geodatabase (in SQL), which is big, complex, has lots of legacy stuff in it that I just don't feel like dealing with right now. Replicating or mirroring the entire database is not an option....because.....in a brilliant management decision, the "datacenter" was placed in a building very far from any modern telco hubs. On a good day I can actually download email attachments. The problem is, the users of this database are every where else but in this building.

More drama: We're drinking the Portal for ArcGIS Cool Aid and high on Collector for ArcGIS. Can't really connect my database to either of those technologies using carrier pigeons.

But I do have some very limited in capacity database and application servers in the "cloud". Certainly not the kind that can house a 7GB water quality database, but here's an idea. What if I could replicate the feature classes that are most needed operationally to these cloud platforms? Field users could get to their data, and I could consume it in my remote datacenter in other applications. All they need access to is the "dots on the map" for their mobile devices, desktop applications, and management access to web-based maps (what stream did that oil spill impact?).

Here's what we'll need:

1. An obnoxiously massive and complex SDE/SQL database called "FISH";
2. A feature class that's been around forever and has a ton of data called "TEST";
3. A remote SDE/SQL database called "FISHREPL" which is the same SQL/SDE version as "FISH";
4. A horrible internet connection.

This is a multi-part blog post, here are the steps:

Real-time Geodatabase Replication? Part 1

Real-time Geodatabase Replication? Part 2

Real-time Geodatabase Replication? Part 3

Real-time Geodatabase Replication? Part 4

Real-time Geodatabase Replication? Part 5

Real-time Geodatabase Replication? Part 6

And of course a few caveats: This only works with unversioned data which places quite a limit on things. In some cases the cons of having unversioned feature classes are not as great as the pros of having a real-time replication service.

Please bear in mind that this method of replication is extremely unsupported. I just did it because I like tinkering under the hood of SDE and wanted to see if it would work. It "sort of" does! However in contrast to the out-of-the-box ESRI replication tools, this process sure resolves a lot of headaches for me. I can attach user-editors to a local SDE instance where "bulk" editing occurs and SQL Merge Replication will fire those changes to the "cloud" SDE instance where the rest of my organization can access (and edit) the data through Portal for ArcGIS, and my mobile clients can perform edits and updates using Collector for ArcGIS. And it works continuously. Full-time. Even with file attachments-enabled and 2 mb photos being attached to a point those changes sync within 60 seconds. SQL replication in 2012 is very efficient in compressing data and ONLY sending deltas. When I generate a "snapshot" of a feature class with file attachments, SQL sees 50 mb of data that needs to be replicated. Performing a versioned-checkout of the same data using the Distributed Geodatabase tools results in a 220 mb file geodatabase.

But emphasis on "experimental". There is a lot of overhead to managing SQL Merge Replication and you can really screw up some operational data. If you get it working after following this blog series I suggest you spend some time deliberately breaking things and see if you can recover your data integrity! Otherwise stick to the ESRI replication tools. They work. When you push the button.....

This is a personal blog and does not recommend, endorse, or support the methods described above. Alteration of data using SQL outside of the ESRI software stack, of course, is not supported and should not be applied to a production database without a thorough understanding and disaster recovery plan.