I'm interacting with an existing database, created in 10.2.0 & recently upgraded to 10.6.1. It contains table entries that participate in a relationship class (M:M). I've noticed that ArcMap cant reliably "show related records" for relationship classes built pre-10.5. Has anyone else experienced this issue? Has there been a fundamental change in the structure of relationship classes in the last several versions? This is not an extended relationship class.
I can confirm having the same issue since upgrading to ArcGIS Desktop 10.6.1 from 10.6
It affects any new FGDB and Relationshipclass created in ArcCat 10.6.1
Origin object class: CAD_GEO_ROW_DBSTN
Destination object class: StreetSegments20190214
Forward Path Label: StreetSegments
Backward Path Label: StreetNames
Message propagation: None (no messages propagated)
Cardinality: 1 - M
Has attributes: No
Origin Primary Key: STREETID
Origin Foreign Key: STREETID
Selecting all street names in CAD_GEO_ROW_DBSTN, and pushing the selection through the relationshipclass, selects all StreetSegments20190214 correctly.
Selecting all StreetSegments20190214, and pushing the selection through the relationshipclass, does NOT select the correct street names in CAD_GEO_ROW_DBSTN. Only 22 of 3227 street names are selected. About 3210 should be selected. This has been going on every time I've created the relationshipclass between the FGDB featureclass and standalone table in a new FGDB since the 10.6.1 upgrade
Not necessarily, there is apparently an additional bug that I was unaware of that impacts queries of more than 99 related objects. The tests above done on record counts less than 99. I will cross post this bug over there, because it is a variable in THEIR testing that they appear to be unaware of.
Hello, we noticed in 10.7 an issue with related records. Our parcels, 1 to many, when selected, we have less records picked (when it should be more) and then when selecting the parcel table to go back, the selection changes and the parcels are not what was originally selected. Is this related to this bug? Thanks
I think this is a bug that can date back as far as the 10.3. In a previous position, I had a ticket open with esri but off the top of my head I can't remember the bug ID. I think it has something to do with what version your database was originally created, and then the upgrading of the database (and indexes) and how the tool functions. But without knowing the inner gears of how the tool works, it's a complete guess. They don't seem to be in a hurry to fix the issue. It's kind of catastrophic in my mind... how many people actually pay close attention to the returned related records? How many false assumptions and analysis have been performed?
I'm with a new organization, just upgraded to 10.6.1 from 10.2.1 and have found this error to be persistant here. I had hoped it was going to be isolated to the previous dataset. Unfortunately.... I'll be opening another support ticket.
Thanks for the response. Yeah, we were getting some crazy results returned and had thought we had corrupted our database in our latest upgrade. As of now, we are just exporting the features classes and tables out of the SQL database and joining them back up with some other software (i.e., Access, Excel, etc.) to get the information we our looking for.
There is a patch available by contacting your esri tech support team.
ArcGIS Desktop 10.6.1 Hot Fix
Hot Fix Name: DT-1061-HF-000004889
Issues addressed in this Hot Fix:
BUG-000116614: The related table returns an incorrect selection result when more than 99
This is resolved at 10.7.1