Arc GIS 10.1 Set Unique identifier field

[JS1]

Hello,

I wanted to perform a Grouping Analysis in ArcGIS 10.1. When opening the analysis tool and trying to select a unique ID field the entries where all blanck. I therefore thought, that mayber the unique identifier field wasn't set. Searching the Arc GIS Help Ressources I found this site that tries to explain how to set the Unique Identifier.
When trying to follow the suggested steps i couldn't set the field because after opening the "source-tab" I could not find anything similar to the mentioned "Change Query" field. The only thing I found is the "set data source" field.
Afterwards i checked all the fields in the table and I found the "FID" in the first collumn. Isn't this field thought to be such a Unique ID field?
Can anyone help me how to manually set this field so I can perform the grouping analysis.
Thx in advance!
Best,
Johannes

[angelagaylard]

Hi Johannes

More than a year later, but I'm having the same problem and suspect I'm doing something silly. Did you ever resolve this issue, and if so, how?

Thanks
Angela

Hello,

I wanted to perform a Grouping Analysis in ArcGIS 10.1. When opening the analysis tool and trying to select a unique ID field the entries where all blanck. I therefore thought, that mayber the unique identifier field wasn't set. Searching the Arc GIS Help Ressources I found this site that tries to explain how to set the Unique Identifier.
When trying to follow the suggested steps i couldn't set the field because after opening the "source-tab" I could not find anything similar to the mentioned "Change Query" field. The only thing I found is the "set data source" field.
Afterwards i checked all the fields in the table and I found the "FID" in the first collumn. Isn't this field thought to be such a Unique ID field?
Can anyone help me how to manually set this field so I can perform the grouping analysis.
Thx in advance!
Best,
Johannes

[JimCousins]

Angela,
Add a field of type Integer, and calculate it equal to the FID. This field will be accessible as the Unique ID Field.
Best Regards,
Jim

[JS1]

Hi Johannes

More than a year later, but I'm having the same problem and suspect I'm doing something silly. Did you ever resolve this issue, and if so, how?

Thanks
Angela

Dear Angela,

I'm sorry to dissapoint you, but I haven't used the tool in the end. Besides being somehow buggy (the tool was only implemented in 10.1) there is no way to interprate the results since ESRI doesn't tell in the tools description which algorithm and distance measurements is in use (see my post here).

From a scientific point of view this is insufficient (at least how it is at the moment). You can produce nice maps but hey: what would have happend if you had used a different distance measurement or altered the algorithm? You can not interpretate your results properly although they might seem to make a nice (presentable) map.

Generally I would recommend you to do the follwing steps if you wan't to do spatial clustering with polygon data in a more reliable way:

Perform normal statistical cluster analyses with different distance measurements and cluster algorithms and then plot your results to different maps (You can use R for this!). From this you will already see if you have somehow regional patterns and you will get a deeper understanding for the nature of your data. Play arround with the variables included, distance measurements, cluster algorithms etc.. Use literature on statistical cluster analysis if you don't have former expirience.

Compare your different Clusterings and try to decide what makes most sense from a formal and theoretical point of view (or your empirical knowledge about the regions). Then look up the maps that represent the most adequate clusterings. If you already have spatial patterns in there, great! Go to step 3! If you don't have spatial patterns... Stop your analysis! Don't use spatial constraint clustering algorithms because they will produce nice regions but meaningless results. 

If you have somehow spatial patterns then ask yourself the following: Do I need spatial contigous regions or not? If not, just leave your results as they are and try to interpret them. If yes, it gets more tricky. Now you have theoretically three options. First: Rearrange your clusters-objects manually (if there are only few polygons that don't match, that procedure is okay). Second: Bring spatial variables into your statistical clustering (e.g. polygon centroids). This will produce more homogenous regions but results get harder to be interpretated. Third: Use tools especially desgined for the purpose of spatial constraint clustering. There is a whole bunch of literature and freeware out there to be used. Google for Regionalization or Spatial Constraint Clustering.

Hope this helps somehow.
Best,
Johannes

[angelagaylard]

Thanks Jim! I also got a similar response from Lauren Scott at ESRI, literally as I pushed the send button!

[CherylCollins]

This was asked long ago, but I just ran across the same problem.  My fix is very roundabout, but it did work.  It only works for point feature classes and assumes that you want to use sequential integers starting at 1, but just in a different order than the ObjectID you have already.

1. Use the Add XY Coordinate tool to add coordinates to your feature class if it doesn't have it already.
2. Select all features and copy them.
3. Open a new spreadsheet and paste the whole table.
4. Delete the existing ObjectID (and Shape) columns.
5. Move the field you want to use as ObjectID to Column A and sort it.
6. Save the table as a text (tab delimited) table (others will probably work as well).
7. In ArcCatalog, right-click the table and select Create Feature Class.
8. Use the Point_X and Point_Y fields for your coordinates.

After I did this I had a new point feature class with the ObjectID exactly in the order I wanted to use.  I hope this helps someone else.

Cheryl