Traits / Segments Calculation Scheduling
|
Collect
1. Core Concepts
1.1. Version (partition)
Version refers to a calculation result for a specific baseline time corresponding to a tag/cluster. Currently, tag/cluster versions are stored daily, with only one version stored per day.
1.2. Computing Basetime (basetime)
Computing Basetime is an identifier for a tag/cluster version and also represents the time benchmark for that version's calculation.
The current baseline time is uniformly taken as 00:00:00 daily. For example, a version generated on 2023-02-22 will have a baseline time of 2023-02-22 00:00:00.
1.3. Update
Update refers to generating a new version based on a new baseline time. Tags/clusters support two update methods: manual updates and routine updates.
- Manual Update: After saving the tag/cluster, a version is generated, and calculation begins immediately. Once the calculation is completed, the data is not updated automatically. You need to manually click update to calculate a new version of the tag/cluster.
- Routine Update: After saving the tag/cluster, a version is immediately calculated. It is subsequently updated according to the update frequency and earliest calculation time configured in the system.
1.4. Recalculate
Recalculate means recalculating data for a historical version based on the original baseline time without generating a new version.
2. Scenario Example
On 2024.02.18 13:00, an operator created a cluster with the rule of including users who performed login events in the past 7 days. This cluster is routinely updated at 02:00 every day.
- Upon initial creation, the system immediately generates a version identified by 2024.02.18 00:00:00. The calculation is based on a 2024.02.18 00:00:00 time benchmark. The calculation range for the past 7 days is 2024.02.11 00:00:00 ~ 2024.02.17 23:59:59.
- When 2024.02.19 02:00 reaches the specified routine update time, the system will automatically start calculating the new cluster version. The id of the new version is 2024.02.19 00:00:00, The time base is 2024.02.19 00:00:00 . The calculation range of the last 7 days 2024.02.12 00:00:00 to 2024.02.18 23:59:59
If the label/cluster rule contains an event that occurred today, the timing of the calculation is different. Example: The operator created a cluster at 2024.02.18 13:00. The classification rules are Users who have logged in today
- If the cluster is updated regularly, the system still uses 2024.02.18 00:00:00 as the time base. Today The corresponding calculation range should be 2024.02.18 00:00:00 to 2024.02.18 23:59:59. However, because the day is not over yet, the system will wait until 2024.02.19 in the morning
- If the cluster is manually updated, it will be calculated immediately after the cluster is created. The calculation range is 2024.02.18 00:00:00 to date All data generated
Note: The content of this document is a technical document that provides details on how to use the Sensors product and does not include sales terms; the specific content of enterprise procurement products and technical services shall be subject to the commercial procurement contract.
