Oliver Ng
11/29/2024, 1:08 PMbucket_start is set in anomalies detection, please?
originally, i thought it'd be set by training_period, however it seems bucket_start is always set to some 100 days before current_date() despite i have configured training_period to be 14 days
appreciate the help ๐๐ผOliver Ng
11/29/2024, 1:11 PMselect edr_bucket_start, edr_bucket_end
from (
select edr_bucket_start,
timestamp_add(cast(edr_bucket_start as timestamp), INTERVAL cast(1 as INT64) day) as edr_bucket_end
from unnest(generate_timestamp_array(cast(
timestamp_trunc(cast(cast('2024-08-23 00:00:00' as timestamp) as timestamp), day)
as timestamp), cast('2024-11-29 12:17:28.562309+00:00' as timestamp), interval 1 day)
) as edr_bucket_start
where timestamp_add(cast(edr_bucket_start as timestamp), INTERVAL cast(1 as INT64) day) <= cast('2024-11-29 12:17:28.562309+00:00' as timestamp)
) results
where edr_bucket_start >= cast(timestamp_trunc(cast(cast('2024-08-23 00:00:00' as timestamp) as timestamp), day) as timestamp)
and edr_bucket_end <= cast('2024-11-29 12:17:28.562309+00:00' as timestamp)
this is how bucket_start and bucket_end are generated and i'd hope to understand what configures bucket_start to 2024-08-23 00:00:00Tiphaine Castel
02/10/2025, 10:56 AM