Hello, I have been playing with Growthbook for a c...
# ask-questions
f
Hello, I have been playing with Growthbook for a couple of days now but I’m still a bit confused about few things, I hope somebody can help me understand: 1. Do I need to have an hypothetical “id-test” feature manually created in the Growthbook UI in order for the growthbook instance running on the client to consider it the same flag when I initialise it with a JSON from a custom source? If yes, do I need the flag name to match or does it need to be fully configured? 2. Does Growthbook look into the data source to identify whether I have run inline tests and produce analysis? And if that’s the case, should I necessarily use the tracking tool as a source in order for the UI to identify tests with data?
If I create this from code, what are the requirements for having “id-test” linked with an “id-test” on the features, and to have the experiments telling me that there is data for an experiment on “id-test” whether I would like to import it?
f
Hi Liborio: 1. No, the SDK does not care how features are defined, and if defined inline, it does not know if they are also created in the GrowthBook app or not. If you call isOn or another SDK method on a feature that is not defined, it will return off or null. 2. No, the data source has no idea how an experiment was run, and does not care.
f
Hi @fresh-football-47124 thanks for the quick reply. I think my message in the thread is more specific.
How is the UI going to determine whether there is new data for an experiment and prompt me to load it? Can this start from a custom source flag?
f
which UI? GrowthBook UI?
f
Correct
Another example is how do I track an inline experiment with the Growthbook experiments UI? I’m assuming Growthbook recognizes its experiments from the data source, but so far my inline tests weren’t automatically tracked.
f
The experiment reports use the data from your data warehouse, typically this is added from the trackingCallback event
it should be easy to debug - either the events are not firing, or they are, and the queries to pull the exposed users is wrong
f
Thanks again, final question I suppose: if I configure the callback correctly, I should be able to initialize an experiment from a feature set in code from a custom json and have the report in the experiments (left alone that I won’t have a way to sync the features displayed in Growthbook UI and without hitting the experiments endpoint of the beta API to post a new experiment, we don’t need that at the moment)
Given that the queries to pull the exposed users are correct, and the trackingCallback is pushing the data correctly, I can set a feature in code and if that contains a test rule, I should be able to find an entry for it in the experiments. Is that correct?
There was an issue with the query, all sorted. Thx