ข้อจำกัดและข้อควรพิจารณาของกระแสข้อมูลDataflows limitations and considerations
มีข้อจำกัดของกระแสข้อมูลบางประการในการเขียน การรีเฟรช และการจัดการความจุที่ผู้ใช้ควรคำนึงถึงดังที่อธิบายไว้ในส่วนต่อไปนี้There are a few dataflow limitations across authoring, refreshes, and capacity management that users should keep in mind, as described in the following sections.
การเขียนกระแสข้อมูลDataflow Authoring
เมื่อเขียนกระแสข้อมูล ผู้ใช้ควรระวังข้อควรพิจารณาดังต่อไปนี้:When authoring dataflows, users should be mindful of the following considerations:
การเขียนในกระแสข้อมูลจะดำเนินการได้ในสภาพแวดล้อม Power Query Online (PQO) ดูข้อจำกัดที่อธิบายไว้ใน ขีดจำกัด Power QueryAuthoring in Dataflows is done in the Power Query Online (PQO) environment; see the limitations described in Power Query limits. เนื่องจากดำเนินการเขียนกระแสข้อมูลในสภาพแวดล้อม Power Query Online (PQO) การอัปเดตที่ดำเนินการบนการกำหนดค่าปริมาณงานกระแสข้อมูลจะส่งผลต่อการรีเฟรชเท่านั้นและจะไม่มีผลกระทบต่อประสบการณ์การเขียนBecause dataflows authoring is done in the Power Query Online (PQO) environment, updates performed on the Dataflows workload configurations only impact refreshes, and will not have an impact on the authoring experience
เฉพาะเจ้าของเท่านั้นที่สามารถแก้ไขกระแสข้อมูลได้Dataflows can only be modified by their owners
กระแสข้อมูลไม่พร้อมใช้งานใน พื้นที่ทำงานของฉันDataflows are not available in My Workspace
กระแสข้อมูลโดยใช้แหล่งข้อมูลเกตเวย์ไม่สนับสนุนข้อมูลประจำตัวหลายรายการสำหรับแหล่งข้อมูลเดียวกันDataflows using gateway data sources do not support multiple credentials for the same data source
การใช้ตัวเชื่อมต่อ Web.Page จำเป็นต้องมีเกตเวย์Using the Web.Page connector requires a gateway
ข้อควรพิจารณาของ APIAPI Considerations
คุณสามารถดูข้อมูลเพิ่มเติมเกี่ยวกับ REST API ของกระแสข้อมูลได้ใน ข้อมูลอ้างอิง REST APIMore about supported Dataflows REST APIs can be found in the REST API reference. ต่อไปนี้คือข้อควรพิจารณาบางประการที่ควรทราบ:Here are some considerations to keep in mind:
การส่งออกและนำเข้ากระแสข้อมูลจะทำให้กระแสข้อมูลนั้นมี ID ใหม่Exporting and Importing a dataflow gives that dataflow a new ID
การนำเข้ากระแสข้อมูลที่ประกอบด้วยเอนทิตีที่เชื่อมโยงจะไม่แก้ไขการอ้างอิงที่มีอยู่ภายในกระแสข้อมูล (ควรแก้ไขคิวรีเหล่านี้ด้วยตนเองก่อนที่จะนำเข้ากระแสข้อมูล)Importing dataflows that contain linked entities will not fix the existing references within the dataflow (these queries should be fixed manually before importing the dataflow)
คุณสามารถเขียนทับกระแสข้อมูลด้วยพารามิเตอร์ CreateOrOverwrite ได้หากสร้างขึ้นในตอนแรกโดยใช้ API นำเข้าDataflows can be overwritten with the CreateOrOverwrite parameter, if they have initially been created using the import API
กระแสข้อมูลอยู่ในความจุแบบแชร์Dataflows in Shared
มีข้อจำกัดสำหรับกระแสข้อมูลในความจุที่แชร์:There are limitations for Dataflows in shared capacities:
- เมื่อรีเฟรชกระแสข้อมูล ระยะหมดเวลาในการแชร์คือ 2 ชั่วโมงต่อเอนทิตีและ 3 ชั่วโมงต่อกระแสข้อมูลWhen refreshing Dataflows, timeouts in Shared are 2 hours per entity, and 3 hours per Dataflow
- ไม่สามารถสร้างเอนทิตีที่เชื่อมโยงในกระแสข้อมูลที่ใช้ร่วมกันได้ แม้ว่าเอนทิตีดังกล่าวจะสามารถอยู่ในกระแสข้อมูลได้ตราบเท่าที่คุณสมบัติ เปิดใช้งานการโหลด บนคิวรีถูกปิดใช้งานLinked entities cannot be created in shared Dataflows, although they can exist within the Dataflow as long as the Load Enabled property on the query is disabled
- ไม่สามารถสร้างเอนทิตีที่คำนวณในกระแสข้อมูลที่แชร์Computed entities cannot be created in shared Dataflows
- บริการ AutoML และ Cognitive ไม่พร้อมใช้งานในกระแสข้อมูลที่แชร์AutoML and Cognitive services are not available in shared Dataflows
- การรีเฟรชแบบเพิ่มหน่วยไม่ทำงานในกระแสข้อมูลที่แชร์Incremental refresh does not work in shared Dataflows
กระแสข้อมูลอยู่ในความจุแบบ PremiumDataflows in Premium
กระแสข้อมูลที่มีอยู่ใน Premium มีขีดจำกัดและข้อควรพิจารณาต่อไปนี้Dataflows that exist in Premium have the following limitations and considerations.
ข้อควรพิจารณาเกี่ยวกับการรีเฟรชและข้อมูล:Refreshes and data considerations:
เมื่อรีเฟรชกระแสข้อมูล ระยะหมดเวลาคือ 24 ชั่วโมง (ไม่มีความแตกต่างสำหรับเอนทิตีและ/หรือกระแสข้อมูล)When refreshing Dataflows, timeouts are 24 hours (no distinction for entities and/or dataflows)
การเปลี่ยนกระแสข้อมูลจากนโยบายการรีเฟรชแบบเพิ่มหน่วยเป็นการรีเฟรชปกติหรือในทางกลับกันจะทำให้ข้อมูลทั้งหมดถูกลบChanging a dataflow from an incremental refresh policy to a normal refresh, or vice versa, will drop all data
การแก้ไขสคีมาของกระแสข้อมูลจะทำให้ข้อมูลทั้งหมดถูกลบModifying a dataflow's schema will drop all data
เอนทิตีที่เชื่อมโยงและคำนวณ:Linked and Computed Entities:
เอนทิตีที่เชื่อมโยงสามารถลงลึกข้อมูลอ้างอิงได้ถึง 32 รายการLinked entities can go down to a depth of 32 references
วงจรอ้างอิงของเเอนทิตี้ที่เชื่อมโยงไม่ได้รับอนุญาตให้ใช้Cyclic dependencies of linked entities are not allowed
คุณไม่สามารถเชื่อมโยงเอนทิตีที่เชื่อมโยงแล้วกับเอนทิตีปกติที่ได้รับข้อมูลจากแหล่งข้อมูลในองค์กรได้A linked entity can't be joined with a regular entity that gets its data from an on-premises data source
เมื่อมีการใช้คิวรี (ตัวอย่างเช่น คิวรี A) ในการคำนวณของคิวรีอื่น (คิวรี B) ในกระแสข้อมูล คิวรี B จะกลายเป็นเอนทิตีที่มีการคำนวณWhen a query (query A, for example) is used in the calculation of another query (query B) in dataflows, query B becomes a calculated entity. คิวรีที่คำนวณไม่สามารถอ้างอิงไปยังแหล่งข้อมูลภายในองค์กรได้Calculated entities cannot refer to on-premises sources.
กลไกการคำนวณ:Compute Engine:
เมื่อคุณใช้กลไกการคำนวณ เวลาในการนำเข้าข้อมูลเพิ่มขึ้นประมาณ 10% ถึง 20% ในช่วงแรกWhile using the Compute engine, there is an approximate 10% to 20% initial increase in time for data ingestion.
- ซึ่งเกิดขึ้นกับกระแสข้อมูลแรกที่อยู่ในกลไกการคำนวณและอ่านข้อมูลจากแหล่งข้อมูลเท่านั้นThis only applied to the first dataflow that is on the compute engine, and reads data from the data source
- กระแสข้อมูลที่ตามมา ซึ่งใช้แหล่งที่มาหนึ่งจะไม่ได้รับโทษเดียวกันSubsequent dataflows, that use the source one will not incur the same penalty
เฉพาะการดำเนินการบางอย่างเท่านั้นที่ใช้กลไกการคำนวณและเฉพาะเมื่อใช้ผ่านเอนทิตีที่เชื่อมโยงหรือเป็นเอนทิตีที่คำนวณOnly certain operations make use of the compute engine, and only when used through a linked entity or as a computed entity. รายการการดำเนินการทั้งหมดมีอยู่ใน บล็อกโพสต์นี้A full list of operations is available in this blog post.
การจัดการความจุ:Capacity Management:
ตามการออกแบบ ความจุ Power BI ระดับพรีเมียมมีตัวจัดการทรัพยากรภายในซึ่งจำกัดปริมาณงานในรูปแบบต่างๆ เมื่อความจุกำลังทำงานบนหน่วยความจำเหลือน้อยBy design, the Premium Power BI Capacities have an internal resource manager which throttles the workloads in different ways when the capacity is running on low memory.
- สำหรับกระแสข้อมูล ความดันการควบคุมนี้จะลดจำนวนคอนเทนเนอร์ M ที่พร้อมใช้งานFor Dataflows, this throttling pressure reduces the number of available M Containers
- คุ๕สามารถตั้งค่าหน่วยความจำสำหรับกระแสข้อมูลเป็น 100% ด้วยคอนเทนเนอร์ที่มีขนาดเหมาะสมสำหรับขนาดข้อมูลของคุณ และปริมาณงานจะจัดการจำนวนคอนเทนเนอร์อย่างเหมาะสมThe memory for Dataflows can be set to 100%, with an appropriately sized container for your data sizes, and the workload will manage the number of containers appropriately
คุณสามารถหาจำนวนคอนเทนเนอร์โดยประมาณได้จากการหารหน่วยความจำทั้งหมดที่จัดสรรให้กับปริมาณงานด้วยจำนวนหน่วยความจำที่จัดสรรให้กับคอนเทนเนอร์The approximate number of containers can be found out by dividing the total memory allocated to the workload by the amount of memory allocated to a container
การใช้กระแสข้อมูลในชุดข้อมูลDataflow usage in datasets
- เมื่อสร้างชุดข้อมูลใน Power BI Desktop แล้วเผยแพร่ไปยังบริการ Power BI ตรวจสอบให้แน่ใจว่าข้อมูลประจำตัวที่ใช้ใน Power BI Desktop สำหรับแหล่งข้อมูล Dataflows เป็นข้อมูลประจำตัวเดียวกับที่ใช้เมื่อชุดข้อมูลถูกเผยแพร่ไปยังบริการWhen creating a dataset in Power BI Desktop, and then publishing it to the Power BI service, ensure the credentials used in Power BI Desktop for the Dataflows data source are the same credentials used when the dataset is published to the service.
- ความล้มเหลวในการตรวจสอบให้แน่ใจว่าข้อมูลประจำตัวเหล่านั้นเป็นผลลัพธ์เดียวกันในข้อผิดพลาด ไม่พบคีย์ เมื่อรีเฟรชชุดข้อมูลFailing to ensure those credentials are the same results in a Key not found error upon dataset refresh
ขั้นตอนถัดไปNext steps
บทความต่อไปนี้ให้ข้อมูลเพิ่มเติมเกี่ยวกับกระแสข้อมูลและ Power BI:The following articles provide more information about dataflows and Power BI:
- ข้อมูลเบื้องต้นเกี่ยวกับกระแสข้อมูลและการเตรียมข้อมูลด้วยตนเองIntroduction to dataflows and self-service data prep
- การสร้างกระแสข้อมูลCreating a dataflow
- กำหนดค่าและใช้กระแสข้อมูลConfigure and consume a dataflow
- การกำหนดค่าที่จัดเก็บกระแสข้อมูลเพื่อใช้ Azure Data Lake Gen 2Configuring Dataflow storage to use Azure Data Lake Gen 2
- ฟีเจอร์พรีเมียมของกระแสข้อมูลPremium features of dataflows
- AI กับกระแสข้อมูลAI with dataflows
- แนวทางปฏิบัติที่ดีที่สุดสำหรับกระแสข้อมูลDataflows best practices