ChuckJonas
Treat duplicate related instances as single reference
Consider the following query:
SELECT Id, Owner.Id, Owner.Name FROM Account
Result:
| AccountId | Owner.Id | Owner.Name |
|---|---|---|
| 001 | 101 | John Doe |
| 002 | 101 | John Doe |
| 003 | 102 | Micheal Jordan |
ts-force will translate EVERY related Owner to a new User instance, even though, Accounts 001 & 002 both have the same 101 Owner.
This is CPU & memory intensive and can potentially lead to a inconsistent state if one object is later mutated:
acc001.owner.Name = 'Jane Doe';
assert(acc001.owner.id, acc002.owner.id); //passes
assert(acc001.owner.name, acc002.owner.name); //fails
We now have the same "instance" in salesforce with two different name values in memory.
However, maybe this IS how it should function? Taking this same idea to the extreme, we would we would track a single instance of an SObject globally, which I'm fairly certain would lead to complicated side effects and is a bad idea...
At the very least, there could be some good performance gains by tracking the different instances during 'mapping' and cloning {...user} instead of running a redundant mapping process. Would be a good intermediate step while I think through the desired behavior.
Maybe even look into using a functional memorized selector (that is GC when the retrieve function returns)
did some testing an surprisingly I didn't really see any performance benefits to doing this... Not sure it's worth it