Anomaly checks when fails

[dinjazelena]

Hey, so when we use anomaly checks which compares DataFrame metrics to previous DataFrame.
Lets say we have batch jobs with pydeequ checks, and one of the checks failed from anomaly check. I go back repair it, but then when i rerun batch job again, it would compare it to failed metric and fail again.

How can i avoid this, or is there option to compare only to baseline DataFrame?

to sum it up:

I have monthly jobs with anomaly checks with lets say relative changes of +-20%, if it fails, job fails, i repair, but then it would compare new run to failed metric and it would fail again.