Indexing（索引（Indexing））¶
In the Ontology, indexing is the process of making tabular or other forms of data in Foundry datasources available for faster data retrieval operations through specialized databases.
This section of documentation describes the indexing process for Object Storage V2, in which indexing is overseen by the Object Data Funnel service ("Funnel"). The Funnel service is responsible for orchestrating Funnel pipelines that create and modify object instances in the Ontology and ensure up-to-date data and metadata.
There are two main types of funnel pipelines, funnel batch pipelines and funnel streaming pipelines, which allow users to adopt one or the other indexing mechanism depending on their datasource landscape, latency and workflow requirements, and cost considerations.
Learn more about Funnel batch pipelines.
Learn more about Funnel streaming pipelines.
For low-latency writes and edits into the Ontology, you can also use direct datasources.
For information about Object Storage V1 (Phonograph) indexing, review the legacy documentation.
中文翻译¶
索引（Indexing）¶
在本体（Ontology）中，索引（indexing） 是指将 Foundry 数据源中的表格或其他形式的数据通过专用数据库进行处理，以实现更快的数据检索操作的过程。
本文档的这一部分描述了对象存储 V2（Object Storage V2）的索引流程，其中索引由对象数据漏斗服务（Object Data Funnel，简称"Funnel"）负责管理。Funnel 服务负责编排 Funnel 管道（Funnel pipelines），这些管道用于创建和修改本体中的对象实例，并确保数据和元数据保持最新。
Funnel 管道主要有两种类型：Funnel 批处理管道（funnel batch pipelines） 和 Funnel 流式管道（funnel streaming pipelines）。用户可以根据自身的数据源环境、延迟和工作流需求以及成本考量，选择其中一种索引机制。
如需对本体进行低延迟写入和编辑，您还可以使用直接数据源（direct datasources）。
有关对象存储 V1（Phonograph）索引的信息，请参阅旧版文档。