How do you stop duplicate data in a feature class from being created?

[TimLangner]

Hi there

I may have asked this before but I'll ask again since 10.1 is now out.

I have a feature class with over 65,000 rows of data. This is stored in ArcSDE 9.3.1, hosted in Oracle 10gr2. I am using ArcMap 10.0.4 myself.

Each row should have a unique feature id. Sometimes features need to be split into two and what I have now found, having done a distinct query on a copy of the feature class in SQL Developer, is 26 rows of duplicated ids.

Is there anyway to stop this from occurring? In Oracle itself I can simply apply a unique key constraint but I've yet to find anything in ArcMap or ArcSDE to stop this from occurring. Is their anything available.

Kind regards

Tim

[VinceAngelo]

You need to handle the split operation for external constraints, of which ArcGIS
has no knowledge. A unique key constraint would prevent you from editing in a
versioned environment, and wouldn't solve the id split issue, so you need either
application code or a trigger (the latter can be risky if you don't handle valid
duplicates [aka versioned edits] correctly).

- V

[TimLangner]

Thank you for your reply vangelo. I'm not looking to do anything risky. I was just looking to do it all within the safe boundaries of ArcGIS itself.

I assume from your comments that it is not possible. I have now upgraded to 10.1 and I get the impression it isn't possible in that either.

Kind regards

[VinceAngelo]

Parcel splitting is a common enough task, but you need to use tools which are aware
of the problem.  A google on 'arcgis parcel splitting' returns what appears to be a
number of ways to address the issue.

- V

[TimLangner]

Thank you for your reply. I had not thought of using any of the land parcel tools. The data set I am working with is a record of features that are maintained. For example grass; flower beds; roads; pavements and so on. I've generally been spiting and merging polygons via digitising the split line for a split or selecting the require polygons and merging them.

Then I would manually change the ID numbers to assign a new one or I would note the unused number and assign it else where to remove id number gaps. I only do this because the process is manual.

Kind regards