The object at location ___ was previously deleted and cannot be retrieved from the Object Stash

Do you still get the same error with feature caching turned off?

Curiously, with the caching turned off the number of errors increases to 80.  The same type of error too.

Hi ​@doug_herrick 

Thanks for the question! I’m still investigating and trying to get a solution but here are some thoughts:

Inspect FeatureJoiner Inputs:

• Ensure both input datasets to the FeatureJoiner have all required features.
• Confirm the join keys are consistent and there’s no filtering or splitting upstream that could lead to feature loss.
• Test with smaller data - verify that the issue isn’t data-related.

Use FeatureHolders:

• Place FeatureHolders directly before the FeatureJoiner inputs and not earlier in the flow.
• Use them only on the datasets being joined, ensuring that features are cached before being consumed by the FeatureJoiner.

If the Problem insists - Replace FeatureJoiner with an Alternative:

• Sorter: Sort input datasets by the join key.
• Joiner: Use the Joiner transformer for controlled feature merging.

I’ll continue to look into this and try to get to the bottom of it. If possible would you be able to share your workspace - this may help with the investigation (though no problem if you can’t).

Let me know if some of these solutions work.
Thank you!
Donal. 

Hi ​@donalmateer,

I didn’t understand all your suggestions (listed above) but I was able to simplify the FME script to the minimum that would still produce the “Object Stash” errors.  It appears to be the data coming out of FeatureJoiner_12 that is causing the errors but I can’t see why.

I’m not able to share the script but I’ve attached the FFS files and here is the setup for the 2 FeatureJoiners.  Hopefully you can reproduce the errors.

Thanks for your time and Happy Christmas.

Doug