Hello all,
I am struggling to get freshness anomal...
# community-supportr
Ryan Silva
04/22/2025, 8:10 PMHello all,
I am struggling to get freshness anomalies working on my dbt models. I created a test dataset designed to fail. The max date is 3 days ago and there are 10 rows for each day. It looks like this:
Ryan Silva
04/22/2025, 8:11 PMWhen I run my tests they succeed when I believe they should fail. Here is my yml file:
Copy code
- name: elementary_test_data
tests:
- elementary.freshness_anomalies:
#anomaly_direction: both
timestamp_column: time_field
# where_expression: < sql expression >
time_bucket:
period: day # Default day: Buckets data by day and looks at max time without update for each bucket
count: 1
# anomaly_direction: spike # "to alert only on delays"
training_period:
period: day # Default Day
count: 14 # Default 14: lookback n buckets to train on
detection_period:
period: day
count: 1 # 2n
Noya Offer
04/27/2025, 1:04 PMHey @Ryan Silva!
It seems that all the data is from the same day, so there's only one time bucket.
Can you try it with 14 days back as you configured? 🙂
r
Ryan Silva
04/27/2025, 2:15 PMThe picture is just the first 10 rows sorted desc. There are actually 20 days in the data, 10 rows each.
n
Noya Offer
04/27/2025, 2:16 PMGot it, apologize for the confusion! Let me take another look
Noya Offer
04/27/2025, 2:17 PMAre you using the UI to visualize the results? Seeing the graph would be very helpful to understand why the test isn't failing
r
Ryan Silva
04/27/2025, 2:19 PMNo, we wanted to just use the cli as a proof of concept before digging in further. Is there a way to see this unpaid?
n
Noya Offer
04/27/2025, 2:21 PMAbsolutely, I'm referring to the static HTML report generated on
edr reportr
Ryan Silva
04/27/2025, 2:21 PMIm actually on my phone, not at the work computer. Can I send that here tomorrow?
n
Noya Offer
04/27/2025, 2:22 PMOf course 🙂
r
Ryan Silva
04/27/2025, 2:22 PMThank you for your help! I appreciate it
🙏 1
7 Views