The issue arises becausetraditional versioned datastores edits indelta tables (Adds and Deletes)instead of the base table. SQL updates applied directly to the base table bypass the delta tables, resulting in incomplete or inconsistent updates.
1. How Traditional Versioning Affects Updates
* In traditional versioning, edits are recorded in delta tables:
* A_<ObjectID> (Adds): Tracks newly inserted rows.
* D_<ObjectID> (Deletes): Tracks deleted rows.
* When SQL commands are executed directly on the base table, they do not affect the data in the delta tables, which causes the feature class to reflect incomplete updates.
2. Why Not Other Options?
* Nonversioned Feature Class that is Partitioned:
* Partitioning organizes data for performance optimization but does not interfere with SQL commands updating the entire table.
* Archiving Enabled on the Feature Class:
* Archiving tracks historical changes in separate archive tables but does not directly impact SQL commands on the feature class.
Steps to Resolve the Issue:
* For traditional versioned data, use thereconcile and postprocess to update values. This ensures that changes are correctly applied across delta tables and the base table.
* Alternatively, use tools likeArcGIS ProorArcPyto programmatically update data instead of executing direct SQL commands.
References from Esri Documentation and Learning Resources:
* Traditional Versioning Overview
* Delta Tables and Traditional Versioning
Conclusion:
The issue occurs because the data istraditional versioned, and direct SQL commands do not account for the delta tables where edits are stored. Use the reconcile and post workflow or ArcGIS tools to apply updates correctly.