Implications of deletion policies（删除策略的影响）¶
:::callout{theme="warning"} Data Lifetime is a valuable yet highly sensitive data governance tool that can lead to severe consequences if misused. Learn more about the risks of Data Lifetime usage and explore safeguards to protect your organization. :::
Fixed deletion date policy¶
When a fixed deletion date policy is applied to a dataset, all transactions across all branches will be assigned a deletion date. If the policy specifies an additional cutoff date, only the transactions that were committed before that cutoff will have a deletion date assigned; any later transactions will not be assigned a deletion date.
:::callout{theme="neutral"} Data Lifetime will not consider policies from other services (the Retention application, for example) when computing deletion dates for transactions. :::
Keep latest view only policy¶
When a keep latest view only policy is applied to a dataset, certain actions may occur depending on the status of a transaction:
- If the transaction is in the latest view of the dataset, on a branch that is included in the set of policies configured for the policy, then the transaction will not receive an assigned deletion date. A transaction will not be marked for deletion if it is in the latest view on any one policy branch.
- If the transaction is not in the latest view of the dataset, the deletion date assigned to a dataset will correspond to the date of the newest transaction’s creation. This rule applies even if the transaction is in an older view within one of the policy's branches, or if the transaction is in the latest view on a branch not covered by the policy.
When a new transaction is added to a dataset with this policy, the following actions will occur:
- If the transaction is on a branch not covered under the policy, it will be assigned a deletion date that matches the transaction's creation date.
- If the transaction is on a branch covered by the policy and does not create a new view (an updated transaction), it will not receive a deletion date.
- If the transaction is on a branch covered under the policy and creates a new view (a SNAPSHOT transaction), it will not be assigned a deletion date. Instead, the dataset will be re-evaluated to place deletion dates on transactions in the previous view(s).
Similar to other policies, all downstream datasets and transactions will inherit the deletion date or its absence from their parent transactions.
中文翻译¶
删除策略的影响¶
:::callout{theme="warning"} 数据生命周期（Data Lifetime）是一项有价值但高度敏感的数据治理工具，若使用不当可能导致严重后果。了解更多关于数据生命周期使用的风险及保护组织的保障措施。 :::
固定删除日期策略（Fixed deletion date policy）¶
当对数据集应用固定删除日期策略时，所有分支上的所有事务（transaction）都将被分配一个删除日期。如果策略指定了额外的截止日期（cutoff date），则只有在该截止日期之前提交的事务才会被分配删除日期；此后的任何事务将不会被分配删除日期。
:::callout{theme="neutral"} 在计算事务的删除日期时，数据生命周期不会考虑其他服务（例如保留应用程序（Retention application））的策略。 :::
仅保留最新视图策略（Keep latest view only policy）¶
当对数据集应用仅保留最新视图策略时，根据事务的状态可能会触发某些操作：
- 如果事务位于数据集的最新视图中，且所在分支包含在策略配置的分支集合内，则该事务将不会被分配删除日期。只要事务位于任一策略分支的最新视图中，就不会被标记为删除。
- 如果事务不在数据集的最新视图中，则分配给数据集的删除日期将对应最新事务的创建日期。即使事务位于策略某个分支的旧视图中，或者事务位于策略未覆盖分支的最新视图中，此规则同样适用。
当向应用了此策略的数据集添加新事务时，将执行以下操作：
- 如果事务位于策略未覆盖的分支上，则会被分配一个与其创建日期相同的删除日期。
- 如果事务位于策略覆盖的分支上，且未创建新视图（即更新事务（updated transaction）），则不会获得删除日期。
- 如果事务位于策略覆盖的分支上，并创建了新视图（即快照事务（SNAPSHOT transaction）），则不会分配删除日期。相反，数据集将被重新评估，为先前视图中的事务设置删除日期。
与其他策略类似，所有下游数据集和事务将从其父事务继承删除日期或缺失删除日期的状态。