All Apps and Add-ons

Understanding "by" grouping and anomaly detection


I wanted to make sure I understand an mstats query that has a "by" clause at the end with regard to machine learning toolkit.

| mstats avg(metric) where index="myindex" by dimension_name

When I choose the dimension name and the metric for detecting categorical outliers are the results based on outliers per dimension or compared to all metrics regardless of the dimension? I think and hope it's the former but wanted to clarify.

Thank you in advance.