跳转至

Concepts glossary（概念术语表）

The following list explains key concepts used in the creation of time series in Foundry. We recommend familiarizing yourself with these concepts to better understand how to effectively set up and use time series in your organization.

Time series

Data that consists of one or more sets of timestamp and value pairs. Each pair represents a quantity at a point in time, thus each set of pairs measures a quantity over time; each set requires a unique identifier. For example, a group of machines may each have several sensors, with each sensor recording values at a regular cadence for the machine housing that sensor; each sensor for each machine will need to be uniquely identified.

Time series object type

An object type (for example, Machine) where each object links to one or more sets of timestamp and value pairs via time series properties (TSPs). Object-centric Foundry applications like Quiver, Vertex, and Workshop allow for analysis of TSPs.

Time series object type backing dataset

A dataset required to create a time series object type; each row represents a single object (for example, Machine 123) of this time series object type and has the following schema:

Column Type Description
Primary key String [Required] A primary key for each row
Series ID String [Required] A series ID column for each TSP
\ \ [Optional] Additional information about each object
\ \ [Optional] Additional configuration information for each TSP

Series ID

A foreign key in the time series object type used to fetch values from a time series sync. Typically, the series ID is distinct within the time series object type backing dataset and has a one-to-many relationship with its corresponding rows in the time series sync.

Time series property (TSP)

A property that provides a value that changes over time (for example, a temperature sensor reading). Each series ID column in the time series object type backing dataset maps to one TSP.

Time series sync

The time series sync indexes time series data, which provides values for TSPs, in Foundry's time series database. Each row that is synced represents a TSP's value for a single point in time (for example, Machine 123’s temperature at 12:00:00 on 01/01/2023 is 100*F). All values for a series ID should be contained in the same sync. These values are fetched through their series ID; thus, a single sync can contain all values for multiple series IDs. The sync should be created as a target output type in Pipeline Builder and contain exactly the following columns:

Column Type Description
Series ID String [Required] The series ID for the set of timestamp and value pairs referred to by a TSP; they must match the TSP's series ID.
Timestamp Timestamp, Long [Required] The time at which the quantity is measured.
Value Integer, Float, Double, String [Required] The value of the quantity at the point that it is measured. A String type indicates a Categorical time series; each categorical time series can have at most 10,000 unique variants.

:::callout{theme="warning"} If a categorical time series exceeds the 10,000 unique variant limit, that series ID will error and no longer be accessible in the platform. :::

Qualified series ID

A qualified series ID has the following shape: "{"seriesId":"<replace with series ID>","syncRid":"<replace with sync RID containing this series ID>"}". The value of a time series property that is backed by multiple time series syncs must be a qualified series ID, and it must be formatted as a JSON string with no newlines or spaces. Learn more about setting up a qualified series ID.

Sensor object type

A time series object type with one default TSP that is linked to an object type for which it records time series data. The linked object type is sometimes referred to as the root object type.

Derived series

Derived series are time series that are calculated with a formula that takes time series as inputs and saves them as a template. These templates are authored in the Time Series Catalog or in Quiver and can be referenced from a time series property for use in Ontology-aware applications.

Bound object type

A time series object type from which a given derived series can be resolved. This object type is specified during derived series creation.

Codex template

A codex template is used to store a templatized derived series formula that is executed at runtime when derived series are visualized or used within other calculations.

中文翻译

概念术语表

以下列表解释了在Foundry中创建时间序列（time series）时涉及的关键概念。建议您熟悉这些概念，以便更好地理解如何在组织中有效设置和使用时间序列。

时间序列（Time series）

由一组或多组时间戳（timestamp）与数值（value）对组成的数据。每一对代表某个时间点的数值，因此每组数据对衡量的是某个量随时间的变化；每组数据对需要唯一的标识符。例如，一组机器可能各自装有多个传感器，每个传感器以固定频率记录其所在机器的数值；每台机器的每个传感器都需要唯一标识。

时间序列对象类型（Time series object type）

一种对象类型（例如机器），其中每个对象通过时间序列属性（TSP）链接到一组或多组时间戳与数值对。以对象为中心的Foundry应用（如Quiver、Vertex和Workshop）支持对TSP进行分析。

时间序列对象类型支撑数据集（Time series object type backing dataset）

创建时间序列对象类型所需的数据集；每一行代表该时间序列对象类型中的一个对象（例如机器123），其架构如下：

列 类型 描述
主键（Primary key） String [必填] 每行的主键
序列ID（Series ID） String [必填] 每个TSP的序列ID列
\<列名> \<列类型> [可选] 每个对象的附加信息
\<列名> \<列类型> [可选] 每个TSP的附加配置信息

序列ID（Series ID）

时间序列对象类型中的外键（foreign key），用于从时间序列同步中获取数值。通常，序列ID在时间序列对象类型支撑数据集中是唯一的，并与时间序列同步中的对应行形成一对多关系。

时间序列属性（Time series property，TSP）

一种提供随时间变化数值的属性（例如温度传感器读数）。时间序列对象类型支撑数据集中的每个序列ID列对应一个TSP。

时间序列同步（Time series sync）

时间序列同步将时间序列数据（为TSP提供数值）索引到Foundry的时间序列数据库中。同步的每一行代表某个时间点TSP的数值（例如机器123在2023年1月1日12:00:00的温度为100°F）。同一序列ID的所有数值应包含在同一个同步中。这些数值通过其序列ID获取；因此，单个同步可以包含多个序列ID的所有数值。该同步应在Pipeline Builder中创建为目标输出类型（target output type），且必须包含以下列：

列 类型 描述
序列ID（Series ID） String [必填] TSP所引用的时间戳与数值对集合的序列ID；必须与TSP的序列ID匹配。
时间戳（Timestamp） Timestamp, Long [必填] 测量该数值的时间点。
数值（Value） Integer, Float, Double, String [必填] 测量时刻的数值。String类型表示分类时间序列（Categorical time series）；每个分类时间序列最多可包含10,000个唯一变体。

:::callout{theme="warning"} 如果分类时间序列超过10,000个唯一变体的限制，该序列ID将报错，且无法再在平台中访问。 :::

限定序列ID（Qualified series ID）

限定序列ID的格式如下："{"seriesId":"<替换为序列ID>","syncRid":"<替换为包含该序列ID的同步RID>"}"。由多个时间序列同步支撑的时间序列属性值必须是限定序列ID，且必须格式化为不含换行符或空格的JSON字符串。了解更多关于设置限定序列ID的信息。

传感器对象类型（Sensor object type）

一种时间序列对象类型，具有一个默认TSP，该TSP链接到为其记录时间序列数据的对象类型。被链接的对象类型有时称为根对象类型（root object type）。

派生序列（Derived series）

派生序列是通过公式计算得出的时间序列，该公式以时间序列为输入并将其保存为模板。这些模板在时间序列目录（Time Series Catalog）或Quiver中创建，可从时间序列属性中引用，用于支持Ontology的应用。

绑定对象类型（Bound object type）

一种时间序列对象类型，从中可以解析出给定的派生序列。该对象类型在创建派生序列时指定。

Codex模板（Codex template）

Codex模板用于存储模板化的派生序列公式，该公式在派生序列被可视化或在其他计算中使用时在运行时执行。