Getting Error "ObjectVersionChanged" after fetching new copy

Ask the Community
, ,
1

Hello,

I have a code repo function that retreives an object, calls my own api to perform an action (which seems like the hack that other people on here are doing), and then updates the object.

This generated an ObjectVersionChanged error.

image
image542×125 6.06 KB

ad9b95f5-8dc5-4a28-bc9d-a27413669d76 if any palantir folks want to look it up.

AIP Assist explains this to be:

The ObjectVersionChanged error typically occurs when attempting to apply an Action to an object that has been modified since it was last retrieved. This error is part of Foundry’s concurrency control mechanism to prevent conflicting updates

This makes sense. So, to fix the bug, when I want to update the object, I now refetch it via:

const newlyFetchedObject= await Objects.search().myObject()
            .filter(myObject => myObject.primaryKey_.exactMatch(previouslyFetchedObject.primaryKey_))
            .allAsync();

newlyFetchedObject.myProperty = newValue;

This is still causing an ObjectVersionChanged error to trigger.

I can’t recreate it in debug mode because, per the docs, debug mode shows me the edits that are authored, not committed. This error, I’m guessing, happens on commit.

Am I doing something wrong? Is there some sort of caching going on? Can I force it to get a fresh object reference?

2

This happens due to caching of search() results. There currently isn’t a way to get a fresh copy of the object from within the same function after applying the action.

The suggested workaround here is to split this into two actions, where the second action performs the edits you are making in the function after calling your first action, and invoke them sequentially based on the order in which they need to execute.