https://www.growthbook.io/ logo
m

microscopic-queen-6171

02/01/2022, 5:27 PM
Hi, In case I have two different tables in BigQuery that I would like to use as sources. Would it be possible to add an experiment that works for both sources? Or will I have to create two copies of the same experiment?
f

fresh-football-47124

02/01/2022, 5:30 PM
different tables or different databases?
m

microscopic-queen-6171

02/01/2022, 5:33 PM
Different tables!
f

fresh-football-47124

02/01/2022, 5:35 PM
yes, the queries are quite customizable and can handle joining across tables- basically anything you can query yourself, you can make it work with GrowthBook
m

microscopic-queen-6171

02/01/2022, 5:42 PM
I guess I could union the two tables while following the experiment_id and variation_id format - however each time I query it, it would have to scan a lot more data. As opposed to one bigquery experiment working for two different bigquery sources
f

fresh-football-47124

02/01/2022, 5:44 PM
We would be happy to jump on a call to help you figure out the queries if you like
there are a couple of ways to do it, and some better than others
f

future-teacher-7046

02/01/2022, 6:34 PM
I'm just curious, what are the two tables? Is it one for dev, one for production? Or split by a dimension like country? Or something else?
m

microscopic-queen-6171

02/02/2022, 8:54 AM
One is standard AB-test experiment viewed, and the other contains data from marketing campaigns. So if I’m running a campaign with two different ads, I want to be able to compare important company metrics between the two attributed groups