https://www.growthbook.io/ logo
Join Slack
Powered by
This message was deleted.
# give-feedback
w
This message was deleted.
f
It looks like that error is coming from our python stats engine. So the query is running fine, but there's an error when processing results. We did make a bunch of changes there recently. Some quick searching suggests it happens when there are NaNs or Infinity in the data. I do see a couple rows in the BigQuery response with all zeros which is suspicious. I'll see if I'm able to reproduce the error.
One thing you could test is adding a custom SQL filter under analysis settings to remove smart tv and see if that helps. 
device <> 'smart tv'
c
i'm getting similar errors when using any dimension to cut results. started within the previous few days and hadn't had issues with this prior
h
I'm investigating now, thanks.
Bryan, can you share the query modal where you're seeing this error? The error may not be tied to the first query in the list as it can be produced by any query in the analysis.
p
Not sure if you fixed something today, but I can't reproduce the error anymore
h
No, we haven't fixed it. But the issue was stemming from that one row where device was 
smart tv
. But we have identified the issue and will be addressing it today.
🙌 2
p
OK, thank you
c
@helpful-application-7107 - do you still need the example query?
h
Is it a ratio metric, @curved-dawn-87799?
And are there some rows where the denominator_sum is 0?
c
possibly. we have about 6 metrics we look at, but it doesnt tell me which failed. removing some that wouldnt have a denominator of 0, the dimension segmentation works
h
No worries.
Hey Bryan. Could you rerun those queries now? We just landed what I think should fix it.
c
✅ all set, confirmed no errors now. thank you!
✅ 1
h
Sascha, let us know if you see this error again. We've landed the fix.
🙌 1
p
Thank you. Yes, will let you know
147 Views
Open in Slack
PreviousNext
Powered by