Add Quiver dashboard to a Marketplace product（将 Quiver 仪表盘添加到 Marketplace 产品中）

Use Foundry DevOps to include your Quiver dashboard in Marketplace products for other users to install and reuse. Learn how to create a Marketplace product.

Supported features

Quiver dashboards with the following features are supported:

• Object analytics cards, such as filter object set, set math, and linked object sets.
• Object-based visualizations (charts) such as bar chart, line chart, and heat grid. Vega plots can also be packaged as long as the underlying transform table is supported.
• Time series charts and plots, including events and grouped plots.
• Tables including most pivot tables, and transform tables with transforms such as group by, join, and filter.
• Basic visualizations including numeric cards, string cards, text cards, and date/time cards are supported.

How to package dashboards

1. From the right side of the dashboard settings pane, select the Enable Marketplace templating option.
2. Select the Run validation button in the header to identify potential Marketplace templatization issues, if any.
3. If the validator does not display any errors, the dashboard Publish or Republish button will also save a Marketplace-ready version of the dashboard.
4. Use that dashboard version in a new or existing Marketplace product.

Linter errors

The linter may indicate up to three main classes of errors:

• Unsupported cards: The card is not currently supported for Marketplace templatization.
• Unsupported configuration: The card is supported, but its configuration is not supported (such as seen in the pivot table examples below). Review the error message provided to learn more.
• Unexpected errors: In most cases, resolving the above two items will remove any unexpected errors. If unexpected errors remain, contact Palantir Support.

:::callout{theme="neutral"} Instead of linting each dashboard individually, you can enable Marketplace linting for an entire analysis to run validation automatically as you work. When analysis-wide linting is enabled, a Marketplace validation indicator appears next to the Save button, displaying the current validation status and the number of errors and warnings found. Select the indicator to review validation details.

In the Global settings panel, select the Enable marketplace linting option to turn on Marketplace templating for all dashboards in the analysis and enable the analysis-wide linter. You can still disable templating for an individual dashboard from its dashboard settings. :::

Inputs required for Marketplace installation

If validation is successful, a View inputs option will appear in the header. Selecting it will display a list of object types, object instances, functions, and link types that the dashboard will require when packaged in Marketplace. For object type inputs, hover over the Properties text to the right of each input to see a list of property types that will need to be mapped.

The Dashboards section of the Analysis History dialog displays which dashboard versions have been validated for packaging:

Create an analysis from an installed dashboard

While the underlying Quiver analysis is not packaged alongside the dashboard, you can create a new Quiver analysis from the installed dashboard by selecting Create analysis from dashboard.

The created analysis will contain a copy of this dashboard, all cards visible in the dashboard, and their upstream dependencies.

:::callout{theme="neutral"} The created analysis and its dashboards will be disconnected from the installed dashboard. Modifications to the analysis will not update the installed dashboard, and changes made to the installed dashboard will not be automatically incorporated into the analysis. :::

Feature usage considerations

The feature's capabilities are still being actively developed. Do take note of the following important usage considerations.

Unsupported cards and configurations

Certain cards which derive their configuration from underlying data have limited support when packaging for Marketplace. See the pivot table and materialization sections below for more details.

Time series

Most Foundry time series features are supported in Marketplace products. However, time series measures are not currently supported in Marketplace.

To ensure compatibility with Marketplace, your dashboard must:

• Not use measures capabilities in time series features and transforms.
• Not have measure columns added to the transform table or object set views. These can be removed from the transform table display settings.

If the linter detects any measures are used, the linter will throw an unsupported configuration error, detailing the offending card and/or measure(s).

As an alternative to using measures capabilities, you can directly import a series from its sensor object, which is a supported method.

Pivot tables

The pivot table card is available but has limitations when a transform table is downstream, affecting column renaming and row operations. Ensure your card:

• Contains no column properties, as the columns would have been based on underlying data.
• Have labels configured the input pivot table card for any aggregations to ensure a stable and unique column name.

Additionally, if there is a Selected Object Set card downstream from a pivot table, its selected data will be cleared before packaging.

Materializations

Cards that use Materialization data types can be packaged, but with limitations. Column names remain unchanged during packaging and installation, so they will not be updated to reflect the installed object type. If the dashboard is installed with a different object type than it was packaged with, Materialization cards (and their downstream cards) within the dashboard may fail to load. Specifically, the property type API names or backing dataset column names must match those in the original object type, depending on whether Rename column names to API names is enabled.

---

中文翻译

---

将 Quiver 仪表盘添加到 Marketplace 产品中

使用 Foundry DevOps 将您的 Quiver 仪表盘包含在 Marketplace 产品中，供其他用户安装和复用。了解如何创建 Marketplace 产品。

支持的功能

支持包含以下功能的 Quiver 仪表盘：

• 对象分析卡片（Object analytics cards），例如筛选对象集、集合运算和链接对象集。
• 基于对象的可视化（图表），例如柱状图、折线图和热力图。只要底层转换表受支持，Vega 图也可以打包。
• 时间序列图表和绘图，包括事件和分组绘图。
• 表格，包括大多数数据透视表，以及包含分组、连接和筛选等转换的转换表。
• 基本可视化，包括数字卡片、字符串卡片、文本卡片和日期/时间卡片。

如何打包仪表盘

1. 在仪表盘设置面板的右侧，选择启用 Marketplace 模板化选项。
2. 点击标题中的运行验证按钮，以识别潜在的 Marketplace 模板化问题（如有）。
3. 如果验证器未显示任何错误，仪表盘的发布或重新发布按钮将同时保存一个适用于 Marketplace 的仪表盘版本。
4. 在新建或现有的 Marketplace 产品中使用该仪表盘版本。

Linter 错误

Linter 可能指示最多三类错误：

• 不支持的卡片：该卡片当前不支持 Marketplace 模板化。
• 不支持的配置：卡片本身受支持，但其配置不受支持（如下文数据透视表示例所示）。请查看提供的错误消息以了解更多信息。
• 意外错误：在大多数情况下，解决上述两个问题将消除所有意外错误。如果意外错误仍然存在，请联系 Palantir 支持。

:::callout{theme="neutral"} 您无需逐个检查每个仪表盘，而是可以为整个分析启用 Marketplace 检查，以便在您工作时自动运行验证。当启用全分析检查时，保存按钮旁边会出现一个 Marketplace 验证指示器，显示当前验证状态以及发现的错误和警告数量。选择该指示器可查看验证详情。

在全局设置面板中，选择启用 Marketplace 检查选项，即可为分析中的所有仪表盘启用 Marketplace 模板化，并开启全分析检查器。您仍然可以在单个仪表盘的设置中禁用其模板化。 :::

Marketplace 安装所需的输入

如果验证成功，标题中会出现查看输入选项。选择它将显示仪表盘在打包到 Marketplace 时所需的对象类型、对象实例、函数和链接类型列表。对于对象类型输入，将鼠标悬停在各输入右侧的属性文本上，可查看需要映射的属性类型列表。

分析历史对话框中的仪表盘部分会显示哪些仪表盘版本已通过打包验证：

从已安装的仪表盘创建分析

虽然底层的 Quiver 分析不会随仪表盘一起打包，但您可以通过选择从仪表盘创建分析，从已安装的仪表盘创建一个新的 Quiver 分析。

创建的分析将包含该仪表盘的一个副本、仪表盘中所有可见的卡片及其上游依赖项。

:::callout{theme="neutral"} 创建的分析及其仪表盘将与已安装的仪表盘断开连接。对分析的修改不会更新已安装的仪表盘，对已安装仪表盘的更改也不会自动纳入分析中。 :::

功能使用注意事项

该功能仍在积极开发中。请务必注意以下重要的使用注意事项。

不支持的卡片和配置

某些从底层数据派生配置的卡片在打包到 Marketplace 时支持有限。有关更多详细信息，请参阅下文的数据透视表和物化部分。

时间序列

大多数 Foundry 时间序列功能在 Marketplace 产品中均受支持。但是，时间序列度量（time series measures）目前不受 Marketplace 支持。

为确保与 Marketplace 兼容，您的仪表盘必须：

• 不在时间序列功能和转换中使用度量功能。
• 不在转换表或对象集视图中添加度量列。这些列可以从转换表显示设置中移除。

如果 linter 检测到使用了任何度量，它将抛出一个不支持的配置错误，详细说明违规的卡片和/或度量。

作为使用度量功能的替代方案，您可以直接从其传感器对象（sensor object）导入序列，这是一种受支持的方法。

数据透视表

数据透视表卡片（pivot table card）可用，但当下游存在转换表时存在限制，这会影响列重命名和行操作。请确保您的卡片：

• 不包含列属性，因为这些列原本基于底层数据。
• 为输入数据透视表卡片配置标签，用于任何聚合操作，以确保列名稳定且唯一。

此外，如果数据透视表下游存在“选定对象集”卡片，其选定数据将在打包前被清除。

物化

使用物化（Materialization）数据类型的卡片可以打包，但存在限制。列名在打包和安装过程中保持不变，因此不会更新以反映已安装的对象类型。如果仪表盘安装时使用的对象类型与打包时不同，仪表盘内的物化卡片（及其下游卡片）可能无法加载。具体来说，属性类型 API 名称或底层数据集列名必须与原始对象类型中的名称匹配，具体取决于是否启用了“将列名重命名为 API 名称”。