This message was deleted.

that's something you do not want to do, unless you can afford doing this in the scope of

apply

I think something like:

Output.json_dumps(dictionary, sort_keys=True).apply(lambda json: sha256(json.encode('utf-8')).hexdigest())

I'm fine with the expense, but it looks like the Output.json_dumps() creates a non-deterministic value.. which is weird since the dictionary's resultant values haven't changed..

Have you set "sort_keys=True"? I'd think that should result in deterministic results. Can you share a copy of what the engine thinks the diff is?

I did, I just realized what it was... I was trying to use the output directly on something instead of running yet another apply against it, so it was returning with the "You can't run str On an Output[T]" bogus... I really think that should raise an exception... 🙂 not create surprise rubbish strings that aren't truly the Output value.