BUG-000156902 for ArcGIS Pro

Numéro d’ID de bogue	BUG-000156902
Envoi	March 22, 2023
Dernière modification	June 5, 2024
S’applique à	ArcGIS Pro
Version trouvée	3.1
Système d’exploitation	Windows OS
Version du système d’exploitation	10.0 64 Bit
Version de correction	3.2
Statut	Fixed

Informations supplémentaires

This issue is addressed in ArcGIS Pro 3.2. There is a combination of things causing this problem. The polygon feature class used as the identity features class has about 9100 features that have 1.6 million parts, and many individual polygon parts span a large area, which at the very least causes performance issues. Also, there are over 30 million areas of overlap within the polygon feature class. In short, the identity polygon feature class is much more complex than might be realized at first glance, and the way in which the polygons are constructed raises the question of whether this is really the best way to represent the data. The geometries themselves are not problems in and of themselves. The issue is triggered by the combination of the data complexity and faultiness with the custom spatial reference that is assigned to the input point feature class. The input feature class's spatial reference is what is used for the analysis, which means the identity feature class is projected on the fly from WGS84 to the input feature class's spatial reference, and then all the features from both inputs are processed using the input feature class's spatial reference. Please refer to the workarounds for how to avoid using this problem's custom spatial reference and get things working.

Solution de contournement

For all workarounds, make sure to set arcpy.env.parallelProcessingFactor = 100 to take advantage of parallel processing and improve performance when running the Identity tool.

Workaround 1

For both the input point feature class and the identity polygon feature class, recreate each feature class to repair the input feature class and to convert the identity feature class to be the same as the input. See https://www.esri.com/arcgis-blog/products/arcgis-pro/analytics/geoprocessing-resolution-tolerance-and-hair/ which contains a step-by-step description on how to recreate and input.

For the step that selects the coordinate system, do not use the input feature class as a template. Instead, at the dialog to select a coordinate system for the new feature class, use the search text box to search for 'statistics Canada Lambert' and select Projected Coordinate System/National Grids/Canada/NAD 1983 Statistics Canada Lambert as the coordinate system for the new feature class.

It is highly recommended that all inputs be projected to the same spatial reference prior to analysis to get the best results. Refer to https://pro.arcgis.com/en/pro-app/latest/tool-reference/appendices/spatial-reference-and-geoprocessing.htm for more information. This workaround takes 3 hours. For faster performance, after converting both feature classes to NAD 1983 Statistics Canada Lambert as in the workaround, convert the identity polygon feature class to a single part using the Multipart To Singlepart tool. Run the Identity tool with the repaired point input and the single part polygon identity feature class. Divide the data into smaller increments and run the tool.