2020 年 1 月January 2020

这些功能和 Azure Databricks 平台的改进已于 2020 年 1 月发布。These features and Azure Databricks platform improvements were released in January 2020.

备注

发布分阶段进行。Releases are staged. 在初始发布日期后,可能最长需要等待一周,你的 Azure Databricks 帐户才会更新。Your Azure Databricks account may not be updated until up to a week after the initial release date.

本月发布了 Azure Databricks 平台版本 3.9 和 3.11。This month saw the release of Azure Databricks platform versions 3.9 and 3.11. 没有发布版本 3.10 或 3.8。There was no release of versions 3.10 or 3.8. 版本 3.7 是只与稳定性和修补程序有关的版本。Version 3.7 was a stability and bug-fix-only release.

即将推出:工作区、池和群集标记将传播到 DBU 使用情况详细信息和 Azure VM,以便改进成本管理报告Coming soon: workspace, pool, and cluster tags propagate to DBU usage details and Azure VMs for better cost management reporting

2 月 10 日,我们将发布 Azure Databricks 使用情况详细信息和 Azure VM 的标记传播。On February 10th, we will release tag propagation to Azure Databricks usage details and Azure VMs. 新的标记传播功能将 Azure Databricks 工作区标记(即资源组标记)、池标记和群集标记合并,并将它们传播到 Databricks DBU 使用情况详细信息和 Azure VM 作为资源标记。The new tag propagation feature combines Azure Databricks workspace tags (that is, resource group tags), pool tags, and cluster tags and propagates them to the Databricks DBU usage details and Azure VMs as resource tags. 你将能够在 Azure 成本管理门户和使用情况详细信息导出内容中查看组合标记信息,从而更好地了解 Azure Databricks 使用情况(总拥有成本)以及各业务部门和团队的使用情况。You will be able to see the combined tag information in the Azure Cost Management portal and in usage detail exports, giving you better visibility into Azure Databricks usage (total cost of ownership) and accurate attribution to business units and teams.

Azure Databricks 和 Azure Lighthouse 现在可以在同一订阅中共存Azure Databricks and Azure Lighthouse can now live in the same subscription

2020 年 1 月 29 日January 29, 2020

所有现有的 Azure Databricks 工作区都从使用托管锁迁移到拒绝分配All existing Azure Databricks workspaces have migrated from using Managed Locks to Deny Assignments. 创建的所有新工作区都将具有“拒绝分配”。All new workspaces created will have Deny Assignments. 这不会改变任何现有行为,并且安全级别保持不变。This does not change any existing behavior, and the level of security remains the same. 尽管你可以加入使用 Azure Databricks 的订阅,但管理租户中的用户目前无法在委托订阅上启动 Azure Databricks 工作区。While you can onboard subscriptions that use Azure Databricks, users in the managing tenant can’t launch Azure Databricks workspaces on a delegated subscription at this time.

用于基因组学的 Databricks Runtime 6.3 正式版Databricks Runtime 6.3 for Genomics GA

2020 年 1 月 22 日January 22, 2020

用于基因组学的 Databricks Runtime 6.3 是基于 Databricks Runtime 6.3 构建的。Databricks Runtime 6.3 for Genomics is built on top of Databricks Runtime 6.3. 它包含用于基因组学的 Databricks Runtime 6.2 的许多改进和升级。It includes many improvements and upgrades from Databricks Runtime 6.2 for Genomics.

关键功能包括:The key features are:

  • 支持 Delta 表充当联合基因分型管道的输入Support for Delta tables as input to the joint genotyping pipeline
  • 读取 VCF 时自动进行批注分析Automatic annotation parsing when reading VCFs
  • 改进了多等位基因变体拆分器Improved multiallelic variant splitter
  • 提升了线性和逻辑回归函数的效率Faster linear and logistic regression functions

有关详细信息,请参阅完整的用于基因组学的 Databricks Runtime 6.3(不支持)发行说明。For details, see the complete Databricks Runtime 6.3 for Genomics (Unsupported) release notes.

Databricks Runtime 6.3 ML 正式版Databricks Runtime 6.3 ML GA

2020 年 1 月 22 日January 22, 2020

Databricks Runtime 6.3 ML 的正式发布引入了许多库升级,其中包括:Databricks Runtime 6.3 ML GA brings many library upgrades, including:

  • PyTorch:1.3.0 到 1.3.1PyTorch: 1.3.0 to 1.3.1
  • torchvision:0.4.1 到 0.4.2torchvision: 0.4.1 to 0.4.2
  • MLflow:1.4.0 到 1.5.0MLflow: 1.4.0 to 1.5.0
  • Hyperopt:0.2.1 到 0.2.2Hyperopt: 0.2.1 to 0.2.2

有关详细信息,请参阅完整的 Databricks Runtime 6.3 ML(不支持)发行说明。For details, see the complete Databricks Runtime 6.3 ML (Unsupported) release notes.

Databricks Runtime 6.3 正式版Databricks Runtime 6.3 GA

2020 年 1 月 22 日January 22, 2020

Databricks Runtime 6.3 的正式发布引入了新功能、改进和许多 bug 修补程序。Databricks Runtime 6.3 GA brings new features, improvements, and many bug fixes.

此版本引入了增强的并发性。This release introduces improved concurrency. 关键功能包括:The key features are:

  • 改进了所有 Delta Lake 操作的并发Improved concurrency for all Delta Lake operations
  • 改进了文件压缩支持Improved support for file compaction
  • 改进了仅插入合并的性能Improved performance for insert-only merge

有关详细信息,请参阅完整的 Databricks Runtime 6.3(不支持)发行说明。For details, see the complete Databricks Runtime 6.3 (Unsupported) release notes.

默认情况下已启用增量缓存Delta cache enabled by default

2020 年 1 月 7-14 日:版本 3.9January 7-14, 2020: Version 3.9

对于所有受支持的 Databricks Runtime 版本,Lsv2 系列实例上默认启用 Delta 缓存。The Delta cache is now enabled by default on Lsv2 series instances for all supported Databricks Runtime releases. 请参阅使用增量缓存See Use Delta caching.

群集标准自动缩放步骤现在可配置Cluster standard autoscaling step is now configurable

2020 年 1 月 7-14 日:版本 3.9January 7-14, 2020: Version 3.9

默认情况下,标准自动缩放的第一步会添加 8 个节点。By default the first step of standard autoscaling adds 8 nodes. 现在,你可以在群集 Spark 配置中设置步骤值。Now you can set the step value in the cluster Spark configuration. 请参阅标准自动缩放See Standard autoscaling.

SCIM API 对“获取用户”和“获取组”支持分页(公共预览版)SCIM API supports pagination for Get Users and Get Groups (Public Preview)

2020 年 1 月 7-14 日:版本 3.9January 7-14, 2020: Version 3.9

SCIM API 现在支持“获取用户”和“获取组”的分页。The SCIM API now supports pagination for Get Users and Get Groups. 指定 startIndexcount 查询参数时,SCIM 将返回部分用户/组。When you specify the startIndex and count query parameters, SCIM will return a subset of users/groups. startIndex 参数是第一个结果的从 1 开始的索引。The startIndex parameter is the 1-based index of the first result. count 参数是要返回的用户或组的最大数量。The count parameter is the maximum number of users or groups to return. 这可确保 SCIM 客户端的可缩放性,并简化 Azure Databricks 管理员的 SCIM 调用。This ensures scalability for the SCIM Client and simplifies SCIM calls for Azure Databricks admins. 请参阅 SCIM APISee SCIM API.

文件浏览器泳道宽度增加到 240pxFile browser swimlane widths increased to 240px

2020 年 1 月 7-14 日:版本 3.9January 7-14, 2020: Version 3.9

增加的宽度减少了将鼠标悬停在对象上以查看完整文件名的需要。The increased width reduces the need to mouse over objects to see the full filename.

Databricks Runtime 3.5 LTS 支持结束Databricks Runtime 3.5 LTS support ends

2020 年 1 月 2 日January 2, 2020

对 Databricks Runtime 3.5 LTS(长期支持)的支持在 1 月 2 日结束。Support for Databricks Runtime 3.5 LTS (Long Term Support) ended on January 2. 请参阅 Databricks 运行时支持生命周期See Databricks runtime support lifecycle.